Search

Reprinted with permission from Lippincott Williams & Wilkins, Rydevik B, Editorial Preface: The Bone and Joint Decade 2000–2010 Task Force on Neck Pain and Its Associated Disorders, SPINE, Volume 33, Number 4S, pp S3

During the last decade, bone anchorage of prosthetic limbs has become a realistic alternative to socket prostheses for patients with amputations due to causes other than severe peripheral vascular disease. The theoretical advantages that could be assumed to exist with an artificial limb attached to the bone as an extension of the residual skeleton, were first described more than 20 years ago (1) and could be summarized as follows: a stable attachment of the prosthetic device with elimination of any movements of a socket, better proprioception due to the stable attachment directly to the bone; elimination of skin and pain problems related to a prosthetic socket; and a better potential to control the prosthetic limb. However, it is through recent clinical research that these advantages have become a reality for patients. The first successful clinical applications of patients treated with bone-anchored amputation prostheses was by the use of the method of osseointegration (2). The discovery that implants made of commercially pure titanium could provide a stable anchorage for an implant in bone tissue was made by the Swedish Professor Per-Ingvar Branemark, during the 1950s. This phenomenon was later named osseointegration (3). Since 1965 the method of osseointegration has been in successful clinical practice for dental applications (4,5) and to date there have been more than two million patients who have been treated with dental implants due to edentulousness worldwide. Currently the same method is, for example, also used for treatment with bone-anchored hearing aids, for anchorage of prostheses due to other defects in the head and neck area (6), for finger joint prostheses (7,8), and for thumb amputation prostheses (9). Treatment with an osseointegrated (OI) transfemoral amputation prosthesis was performed for the very first time in 1990 in Sweden (2). Since then, the treatment has continued to be performed in Sweden and more recently also in the United Kingdom (10,11). Until today, most patients treated have had a transfemoral amputation, but individuals with amputations on the upper extremity, that is, transhumeral, transradial, and thumb amputations, have also been treated with OI prostheses. To date, more than 120 patients have been treated with OI prostheses worldwide, with most patients treated at the Sahlgrenska University Hospital in Gothenburg, Sweden. The treatment includes two surgical sessions. At the first surgery (S1), a titanium implant (fixture) is inserted in the residual bone and left unloaded for 3 to 6 months. At the second surgery (S2), a titanium rod (abutment) is inserted into the distal end of the fixture and then penetrates out of the skin on the residual limb (Fig. 10.1). Prosthetic suspension is obtained by connecting the OI prosthesis to the abutment with a specific attachment device (Figs. 10.2 and 10.3). After the second surgery, the patient undergoes a period of rehabilitation. For

Study design Assessment of activated T and B cells in a subcutaneous chamber filled with autologous nucleus pulposus using flow cytometry and immunohistochemistry. Objectives To examine if subcutaneously placed autologous nucleus pulposus may attract activated T and B cells in an animal model. Summary of background data Nucleus pulposus has been suggested to trigger an autoimmune response if exposed to the immune system, for example, in association with disc herniation. T-cell activation represents a hallmark in the generation of an autoimmune response, subsequently leading to the differentiation of B cells, but a causal association between the exposure of nucleus pulposus to the systemic circulation and T and B cell activation is still lacking. Methods Autologous nucleus pulposus was harvested from the intervertebral disc of 9 pigs and placed subcutaneously in perforated titanium chambers. In order to control for the effect of the titanium chamber, an additional empty chamber was placed subcutaneously in each pig. After 7 days, the pigs were killed and the chambers were harvested. Flow cytometry and immunohistochemistry were used for analysis of T-helper cells (CD4+), cytotoxic T cells (CD8+), and B cells (Igkappa) in the chamber exudates and T cells (CD45RC) in the remaining blood clot tissue of the chamber. Results As compared with the empty chambers, the proportion of activated T cells (CD4+ and CD8+) was significantly higher in the exudate of the nucleus pulposus filled chamber. The proportion of activated B cells expressing immunoglobulin kappa (Igkappa) was also significantly elevated in the exudate of the nucleus pulposus chambers. The analysis of the remaining chamber tissue revealed a significantly higher amount of T cells (CD45RC) in the nucleus pulposus chambers than in the empty chambers. Conclusions The present findings indicate that nucleus pulposus attracts activated T and B cells. However, since the cell population in the nucleus pulposus of young pigs may differ from that of adult humans, the obtained data may not be directly transferred to the human situation of a disc herniation. The observations in the present study may nevertheless explain some of the local tissue reactions occurring in association with disc herniation and nerve root involvement, thereby providing further insight into the pathophysiology of sciatica.

Study design The mechanisms behind the formation of a characteristic tissue reaction at the surface of the dorsal root ganglion (DRG) exposed to nucleus pulposus was studied with special reference to apoptosis using electron microscopy and immunohistochemistry in rats. Objectives To study the mechanism of the characteristic tissue reaction at the surface of the DRG exposed to nucleus pulposus. Summary of background data Recently, it was observed that local application of nucleus pulposus may induce a characteristic tissue reaction at the surface of the DRG. This change occurred as early as 1 day after the application of nucleus pulposus. METHODS.: Herniation of nucleus pulposus was created in the L4-L5 disc in rats. The L4 DRG were resected 3 and 24 hours after surgery. The sections of the specimens were observed using light and electron microscopy. The sections were processed for immunohistochemistry using antibodies to single-stranded DNA (ssDNA), caspase 3, and tumor necrosis factor alpha (TNF). Results There were typical changes of the cell nuclei observed by light and electron microscopy, especially those of the small-sized cells, in the DRG 24 hours after application of nucleus pulposus, indicating the presence of apoptosis. The presence of ssDNA, caspase 3, and TNF further enhanced the impression that there was apoptosis in the DRG. Nucleus pulposus induced apoptosis in the DRG at the site of application within as little as 24 hours. Conclusions Nucleus pulposus herniated from the disc induced apoptosis in at the surface of the DRG exposed to nucleus pulpous as early as 24 hours after exposure.

Different fracture fixation techniques and fracture environment influence bone formation in healing fractures. However, the influence on the development of biomechanical properties has not been clear described. We evaluated the influence of fracture fixation stability and fracture environment on mechanical properties in healing femoral fractures in rats.Animals were treated surgically with external fixation: 1 group (27 rats) with 0-mm fracture gap size with bone ends touching, corresponding to an axial stiffness of 265 (SD 34) N/mm, and a second group (27 rats) with 2-mm fracture gap size corresponding to an axial stiffness of 30 (SD 2.1) N/mm. From each group, 6-7 animals were killed at 2, 4, 6 and 12 weeks. Torsional test revealed a delay in torsional stiffness in fractures in group 2 compared to group 1. In group 2, the torsional stiffness of the contralateral femora was found to be greater at 12 weeks than the torsional stiffness in group 1.We found that during fracture healing, the development of torsional stiffness corresponds to the magnitude of endochondral ossification and late response of bone formation. A significantly increased torsional stiffness in the non-fractured leg of rats with delayed fracture healing was also found, possibly indicating a response to loading conditions or a systemic stimulation of bone mass.

STUDY DESIGN: The effect of infliximab, a chimeric monoclonal antibody to TNF-alpha, on induction of brain-derived neurotrophic factor (BDNF) was examined using an experimental herniated nucleus pulposus (NP) model. OBJECTIVES: To investigate whether treatment of infliximab could attenuate an induction of BDNF, which functions as a modulator of pain, following NP application to the nerve root. SUMMARY OF BACKGROUND DATA: Evidence from basic scientific studies proposes that TNF-alpha is involved in the development of NP-induced nerve injuries. However, the therapeutic mechanisms of infliximab against pain have not been elucidated experimentally. METHODS: Twenty rats were used in this study. In the test groups, the animals underwent application of NP to the L4 nerve roots and received a single systemic (intraperitoneal) injection of infliximab at the time of surgery (Infli-0 group, n = 5) or at 1 day after operation (Infli-1 group, n = 5). As a control treatment, sterile water was administered intraperitoneally to 5 rats with NP application (NP group) and to 5 sham-operated rats (sham group). On day 3 after surgery, the L4 dorsal root ganglion (DRG) and L4 spinal segment were harvested and assessed regarding BDNF immunoreactivity. RESULTS.: Application of NP induced a marked increase of BDNF immunoreactivity in number in the DRG neurons and within the superficial layer in the dorsal horn compared with the sham group (P < 0.01). Infliximab treatment in the Infli-0 and Infli-1 groups reduced the BDNF induction in both DRG and spinal cord (P < 0.05). CONCLUSION: These findings indicate that infliximab attenuates the elevated BDNF levels induced by NP. The present study therefore further indicates the importance of TNF-alpha in sciatica due to disc herniation and the possible therapeutic use of a TNF-alpha inhibitor for this condition.

Application of nucleus pulposus to the spinal nerve and displacement of the adjacent nerve results in behavioral changes in rats. It has been reported that treatment with the tumor necrosis factor-alpha (TNFalpha) inhibitor, infliximab, significantly reduces spontaneous pain behavior in this animal model. However, there have been no reports of the effects of infliximab on mechanical or thermal hyperalgesia using this model. Disk incision and adjacent spinal nerve displacement were performed with (n = 6) or without (n = 6) injection of infliximab. A control group also underwent sham surgery (n = 6). The animals were tested for 3 days before and on days 1, 3, 7, 14, and 21 after surgery. Non-noxious mechanical thresholds were tested by determining the hind paw withdrawal response to von Frey hair stimulation of the plantar surface of the footpad with a touch stimulator. Thermal nociceptive thresholds were tested using a sensitive thermal testing device. While disk incision with displacement surgery rats showed mechanical and thermal hyperalgesia after surgery on the experimental side, neither rats treated with infliximab nor the sham operation controls showed these effects. Injection of infliximab seemed to prevent mechanical and thermal hyperalgesia caused by the combination of disk incision and nerve displacement.

STUDY DESIGN To analyze the effects of tumor necrosis factor-alpha, interleukin-1beta, and interferon-gamma on cauda equina function and to define if any of these cytokines could induce nerve root dysfunction comparable with the situation with application of nucleus pulposus. SUMMARY OF BACKGROUND DATA Proinflammatory cytokines derived from the intervertebral disc have been suggested to mediate the nucleus pulposus-induced nerve root injury following local application of nucleus pulposus. However, it is not known if such cytokines may induce similar injury if applied separately. METHODS A total of 29 pigs were used. Nucleus pulposus was harvested from lumbar discs and applied to the sacrococcygeal cauda equina following laminectomy of the first coccygeal vertebra in seven pigs. Five pigs received 1.66 microg of tumor necrosis factor-alpha, five pigs received 0.85 microg of interleukin-1beta, and five pigs received 1.66 microg of interferon-gamma. Seven pigs received autologous fat for control. Nerve conduction velocity was studied by local electrical stimulation and recordings in the tail muscles 7 days after the application. RESULTS Application of nucleus pulposus and fat induced similar effects as seen in previous studies, with normal nerve conduction velocity for fat and a significant reduction for nucleus pulposus. Application of both interleukin-1beta and IFN-gamma induced slight reductions of nerve conduction velocity compared with fat, but they were not statistically significant. Tumor necrosis factor-alpha, however, induced a reduction of the velocity that was even more pronounced than for nucleus pulposus. CONCLUSION Based on previous observations and the data of the present study, one may conclude that tumor necrosis factor-alpha from nucleus pulposus cells seems to be intimately involved with the basic pathophysiologic events leading to both nerve root dysfunction and pain after local, epidural application of nucleus pulposus. One may therefore also suspect that pharmacologic inhibition of tumor necrosis factor-alpha may at least theoretically be considered in the clinical situation with disc herniation and sciatica.

STUDY DESIGN The effects of diclofenac and ketoprofen on nerve conduction velocity in experimental nerve root compression were evaluated in a setup using an established pig model. OBJECTIVE To assess the effects of two potent nonsteroidal antiinflammatory drugs, diclofenac and ketoprofen, in experimental nerve root compression. SUMMARY OF BACKGROUND DATA Compression of spinal nerve roots is recognized to be of major etiologic importance for several common spinal pain syndromes. Secondary inflammatory changes, induced by microvascular permeability changes and leakage of inflammatory mediators into the endoneural tissue, have been proposed as important for the induction of spinal nerve root injury by chronic compression. METHODS This study involved 21 pigs. An ameroid constrictor was used to induce compression. Seven pigs were treated with daily intramuscular injections of diclofenac 3 mg/kg for 7 days. Seven other pigs were treated with daily intramuscular injections of ketoprofen 4 mg/kg. For a control, seven pigs did not receive any drug treatment. After 7 days, the pigs were reanesthetized, and the nerve conduction velocity in the compressed nerve root segments was determined. RESULTS The nerve conduction velocity was significantly higher (P < 0.05, Student's t test) in the pigs treated with diclofenac (50 +/- 16 m/second) than in the untreated pigs (32 +/- 15 m/second). The nerve conduction velocity also was significantly higher (P < 0.05) in the pigs treated with ketoprofen (59 +/- 16 m/second) than in the untreated pigs. There were no significant differences in nerve conduction velocity between pigs treated with ketoprofen and those treated with diclofenac. CONCLUSIONS The findings indicate that intramuscular administration of diclofenac or ketoprofen, both potent antiinflammatory drugs, may reduce nerve root dysfunction induced by compression of spinal nerve roots in an experimental pig model.

Study design Nerve conduction velocity was measured in the pig cauda equina after local application of anulus fibrosus or in vitro/postmortem degenerated nucleus pulposus from the same pig. Objectives To analyze the effects of anulus fibrosus and degenerated nucleus pulposus on nerve conduction velocity. Summary of background data Previous studies on nucleus pulposus-induced effects on nerve roots have used normal, nondegenerated nucleus pulposus. Because both anulus fibrosus and degenerated nucleus pulposus are commonly seen in the clinical situation of disc herniation, the value of the previous work could be questioned. Methods Anulus fibrosus and nucleus pulposus were harvested using a retroperitoneal approach. The nucleus pulposus was degenerated artificially either by addition of sodium lactate with HCl added to form a pH of either 6.0 or 3.5 (in vitro degeneration), or by storing the nucleus pulposus at 4 C until a pH of 6.0 (postmortem degeneration) was reached. After epidural application, the nerve conduction velocity was determined at 7 days (anulus fibrosus) or 3 days (degenerated nucleus pulposus). Results Application of anulus fibrosus did not induce any reduction of nerve conduction velocity. In vitro and postmortem degenerated nucleus pulposus induced a reduction of nerve conduction velocity similar to that of normal nucleus pulposus. Conclusions Although only nerve function and not pain was assessed, it seems likely that previous experiments using normal nucleus pulposus may be relevant for evaluating the pathophysiologic mechanisms behind the nucleus pulposus-induced nerve root injury, also in a clinical perspective.

Main problem. Nucleus pulposus and/or chronic compression can induce spinal nerve root injury. Inflammation has been proposed as having major importance in the pathophysiologic mechanisms involved in the induction of such injuries. Corticosteroids, potent anti-inflammatory drugs, have been demonstrated to reduce nucleus pulposus-induced spinal nerve root injury. The aim of the present study was to assess the effects of two potent nonsteroidal anti-inflammatory drugs (NSAIDs), diclofenac and ketoprofen, in experimental nucleus pulposus-induced spinal nerve root injury in a pig model. Methods. Eighteen pigs were included in the study. Autologous nucleus pulposus was harvested from a lumbar disc and applied locally around the first sacral nerve root after a partial laminectomy of the first and second sacral vertebrae. Six pigs were treated with daily intramuscular injections of diclofenac, 3 mg/kg body weight, for 7 days. Six other pigs were treated with daily intramuscular injections of ketoprofen, 4 mg/kg body weight, for 7 days. As controls, six pigs received injections with physiologic saline. After 7 days, the pigs were reanesthetized and the nerve conduction velocity over the exposed nerve root area was determined. Results. The nerve conduction velocity was significantly higher in pigs treated with diclofenac than in the saline-treated controls, (57±6 m/s vs 38±18 m/s, P

Proinflammatory cytokines have been identified in herniated intervertebral discs in humans, and such cytokines have experimentally been demonstrated to be important in the pathophysiological mechanisms of disc herniation. Cerebrospinal fluid (CSF) and serum concentrations of interleukin (IL)-1beta IL-6, IL-8, interferon (IFN)-gamma and tumour necrosis factor (TNF)-alpha were investigated using the enzyme-linked immunosorbent assay (ELISA) technique in 39 patients with lumbar disc herniation and sciatica. Pain duration and pain intensity (visual analogue scale, VAS) were recorded at inclusion, and a clinical examination was performed evaluating neurological findings. The extent of disc herniation (protrusion or extrusion/sequestration) was evaluated perioperatively. Normal concentrations of IL-1beta, IL-6, IFN-gamma and TNF-alpha were present in CSF and serum in almost all patients with lumbar disc herniation. The concentrations of IL-8 in CSF were increased in 12 out of 39 patients, and these increased levels of IL-8 correlated to a short duration of pain and to more pronounced herniation (extrusion or sequestration). No relationship between IL-8 concentrations in CSF and pain intensity, positive neurological findings or a positive straight leg-raising (SLR) test was found. The observation of increased concentrations of IL-8 in CSF in patients with a short duration of symptoms supports the concept of the initial involvement of inflammatory mechanisms after a disc herniation. The finding that most of the patients with increased concentrations of IL-8 in CSF had an extrusion or a sequestration may suggest that the increase in IL-8 is related to mechanical nerve root compression, but may also indicate a biochemical effect exerted by the herniated disc on the surrounding tissue. Further studies on the potential role of IL-8 as a biomarker for disc herniation are warranted.

STUDY DESIGN The possibility to prevent nucleus pulposus-induced functional and structural nerve root injury by selective tumor necrosis factor-alpha inhibition was assessed in an experimental model in the pig spine. OBJECTIVE The objective of the study was to evaluate the role of tumor necrosis factor-alpha in the mediation of nucleus pulposus-induced nerve injury by using selective inhibition. SUMMARY OF BACKGROUND DATA The cytokine tumor necrosis factor-alpha has been suggested to play a key role in the nerve root injury induced by local application of nucleus pulposus. However, previous studies have not been able to distinguish the effects between tumor necrosis factor-alpha and other disc-related cytokines because of the use of nonspecific cytokine inhibition. METHODS Autologous nucleus pulposus was harvested from a lumbar disc and applied to the porcine sacrococcygeal cauda equina. The pigs were simultaneously treated with two selective tumor necrosis factor-alpha inhibitors (etanercept n = 8 and infliximab n = 5), a heparin analogue (enoxaparin n = 5) or saline for control (n = 5). After 7 days the nerve conduction velocity over the application zone was determined and samples of the exposed nerve roots were collected for light microscopic evaluation. RESULTS The two tumor necrosis factor-alpha inhibitors prevented the reduction of nerve conduction velocity and also seemed to limit the nerve fiber injury, the intracapillary thrombus formation, and the intraneural edema formation. However, treatment with enoxaparin did not seem to be different from control regarding reduction of nerve conduction velocity or histologic changes. CONCLUSIONS The data clearly indicate that tumor necrosis factor-alpha is involved in the basic pathophysiologic events leading to nerve root structural and functional changes after local application of nucleus pulposus. The study therefore provides a basic scientific platform with potential clinical implications regarding the use of anti-tumor necrosis factor-alpha medication as treatment in patients with disc herniation and sciatica.

Limited data are available about the long-term outcome of surgical treatment for lumbar spinal stenosis, and there is a wide variation in reported success rates. There is also a controversy regarding differences in long-term outcome between patients undergoing decompressive surgery alone and those undergoing both decompression and fusion. The aim of this study was to evaluate the long-term clinical outcome and possible complications of decompressive surgery, with special reference to possible differences between patients undergoing fusion, with or without instrumentation, and those undergoing decompression alone. All 124 patients undergoing first-time surgery for lumbar spinal stenosis between 1982 and 1991 at our department were included, and their medical records were reviewed retrospectively. Ninety-six of the patients were available for follow-up and were re-examined by an independent investigator and assessed with a questionnaire after a mean follow-up period of 7.1 (range 4.0–12.2) years. Sixty-five percent of all the patients at the follow-up were subjectively satisfied. Eighty-eight percent of the patients reported constant or daily leg pain preoperatively compared to 43% at follow-up. Constant or daily low back pain was reported by 83% of the patients preoperatively compared to 45% at follow-up. Improvement in walking capacity was found in most patients, and only 4% of the patients who had a preoperatively documented maximum walking distance reported a decreased walking capacity. Twenty-four (25%) of all patients used analgesics daily at the time of follow-up, 34 patients (35%) occasionally and 38 patients (40%) never. The patients with fusions, instrumented or non-instrumented, did not differ significantly from the unfused patients regarding any of the above-mentioned parameters. The results of the study showed that most patients demonstrated a considerable improvement in walking capacity at follow-up. This improvement was significant (P < 0.001) and of clinical importance. A significant improvement regarding both low back pain and leg pain was found postoperatively compared to preoperatively (P < 0.001). There were no statistical differences, judged by all the evaluated parameters, regarding the clinical outcome between patients who were fused and those who were not. Neither were any significant differences found between instrumented fusions compared to uninstrumented fusions. In accordance with most other long-term follow-up studies, about two-thirds (65%) of the patients claimed a satisfactory result at follow-up.

In the present study the psychophysical detection threshold levels for mechanical stimulation of 32 prosthetic limbs were determined. Prosthetic limbs were anchored to the bone by means of an implant (n=17) or supported by a socket enclosing the amputation stump (n=15). Detection threshold levels were assessed for pressure and vibratory stimulation of the prosthesis and the limb at the contralateral side (control). Following vibratory stimulation, thresholds were increased on an avarage 20% for socket prostheses, but approached those of the control for bone-anchored prostheses. For pressure stimulation, thresholds were increased up to 60% for socket prostheses and 40% for bone-anchored prostheses compared to the control. While bone-anchored prostheses yielded significantly lower threshold levels than socket prostheses, there was no significant difference between both treatments regarding pressure stimulation. Results were applicable to both upper and lower limb amputees. It could be concluded that detection thresholds for pressure and especially vibratory stimulation of prosthetic limbs were generally higher than for control limbs. The outcome was related to the prosthetic limb design with bone-anchored prostheses yielding better perception than socket prostheses. All correspondence to be addressed to R. Branemark

Although it is well established that nucleus pulposus cells may induce structural and functional changes in adjacent nerve roots when placed epidurally, it is not known whether this is due to direct neurotoxic effects or whether the nerve roots are affected indirectly by reduction of nutrition and inflammatory/immunologic mechanisms. In the present study we assessed the effects of various tissues on cultured dorsal root ganglions from newborn rats. Nucleus pulposus was found to have a toxic effect on the axons by blocking axonal outgrowth, but no similar effects on the nerve cell bodies (extra-ganglionic nerve cell density, nerve cell arborisation) were found as compared to the series with only culture medium. Sterile water for 1 or 24 h (positive controls) induced significant effects by all four criteria, whereas medium without nerve growth factor, fat and frozen nucleus pulposus had no statistically significant effects. The study thus showed that there are direct axonotoxic effects induced by the nucleus pulposus, and since frozen nucleus pulposus did not have any effects, it may be assumed that the mechanisms are related to substances produced by the nucleus pulposus cells. The presented model allows for future studies on the neurotoxic properties of nucleus pulposus cell-derived candidate substances.

STUDY DESIGN Measurement of changes in cerebrospinal fluid concentrations of nerve tissue markers, total proteins, and immunoglobulin after compression of nerve root or application of nucleus pulposus in a pig model. OBJECTIVES To assess whether compression or application of nucleus pulposus to spinal nerve roots may cause increased levels of cerebrospinal fluid markers of nerve tissue damage and total proteins, and whether synthesis of immunoglobulins may be induced in cerebrospinal fluid. SUMMARY OF BACKGROUND DATA Previous studies have reported that there seems to be a relationship between elevated cerebrospinal fluid total protein concentrations, nerve tissue markers, clinical findings, and compression of the nerve root evaluated by radiologic changes in patients with sciatica. METHODS Subjects included 41 pigs, including 5 control animals. In two groups of experimental animals (n = 7; n = 5), an ameroid constrictor was slid onto the S1 nerve root. In two other groups (n = 7; n = 5), nucleus pulposus harvested from the L2-L3 disc was applied to the S1 nerve root. Two sham animal groups (n = 7; n = 5) underwent the same laminectomy. Twenty-one pigs underwent reoperation after 1 week, and 15 pigs after 4 weeks. A syringe was used to remove 3 mL of cerebrospinal fluid at L4-L5. Concentrations of total proteins, the light subunit of the neurofilament protein, S-100 protein, neuron-specific enolase, and glial fibrillary acidic protein were measured, and the presence of oligoclonal bands (immunoglobulins) were assayed in cerebrospinal fluid. RESULTS The pigs with compressed S1 nerve root had considerably higher neurofilament protein and total protein concentrations in their cerebrospinal fluid than the-control animals (P < 0.001 and P < 0.01, respectively) or the sham animals (P < 0.001 and P < 0.05) in the 1-week experiment. Nucleus pulposus did not induce a significant increase in concentrations of the different protein markers. The presence of oligoclonal bands in cerebrospinal fluid in the experimental groups did not differ between the control and sham animals. CONCLUSIONS The neurofilament protein and total protein concentrations in cerebrospinal fluid may have diagnostic importance in cases wherein clinical findings are not clearly related to the radiologic changes and vice versa. These protein markers also may be useful tools in different experimental models.

STUDY DESIGN The light subunit of neurofilament protein, S-100 protein, neuron-specific enolase, and glial fibrillary acidic protein were determined in the cerebrospinal fluid in patients with lumbar disc herniation and in control patients. OBJECTIVES To determine whether nerve root injury caused by disc herniation increases the levels of nerve and glial cell injury markers in the cerebrospinal fluid. SUMMARY OF BACKGROUND DATA Markers of nerve tissue injury can be analyzed in the cerebrospinal fluid, allowing characterization of the cell types involved and the degree of disease in patients with neurologic disorders. METHODS Cerebrospinal fluid samples were obtained by preoperative lumbar puncture in patients who underwent surgery for lumbar disc herniation and in patients who underwent lower extremity surgery (control group), neurofilament protein (light subunit) and glial fibrillary acidic protein were analyzed by enzyme-linked immunosorbent assay and S-100 protein and neuron-specific enolase by radioimmunoassay and luminescence immunoassay, respectively. In the disc herniation group the concentrations of the four markers were evaluated regarding possible correlation to patient history, computed tomographic findings, and clinical findings. RESULTS Cerebrospinal fluid concentration of neurofilament protein (light subunit) and S-100 were increased in the disc herniation group compared with that in control subjects (1158 +/- 383 ng/L vs. 152 +/- 14 ng/L, P < 0.01; 1963 +/- 231 ng/L vs. 1003 +/- 152 ng/L, P < 0.05, respectively). No statistical differences in neuron-specific enolase and glial fibrillary acidic protein concentrations were observed between the groups. Disc herniation patients with fewer than 3 months' duration of subjective symptoms had higher neurofilament protein levels than did patients with longer duration. None of the markers was related to preoperative clinical or computed tomographic findings. Patients with persistent neurologic findings at follow-up 2-3 months after surgery had higher levels of neurofilament protein before surgery compared with-those without sequelae. CONCLUSIONS Patients with disc herniation and sciatica have increased concentrations of neurofilament protein and S-100 in the cerebrospinal fluid, which indicates damage of axons and Schwann cells in the affected nerve root.

STUDY DESIGN Compression onset rate, anatomic aspects, and morphologic and neurophysiologic effects in spinal nerve roots were studied in a nerve root compression model in pigs. OBJECTIVES To analyze the compression onset rate by measuring the gradual reduction of the inner diameter of the constrictor, the motor nerve conduction velocity by electromyography, the morphologic changes by light microscopy, and the gross and vascular anatomy by dissection and ink injections, respectively, in a model for experimental chronic nerve root compression. SUMMARY OF BACKGROUND DATA Chronic nerve root compression is recognized to be related to back pain syndromes, including sciatica. Various aspects of morphologic and physiologic changes have been studied previously in models for acute compression and chronic nerve root irritation, but a controlled, graded chronic nerve root compression model has not been described. METHODS An ameroid constrictor was applied around a spinal nerve root just cranial to the dorsal root ganglion. The inner diameter of this constrictor gradually becomes reduced. After 1 week or 4 weeks, electromyographic measurements were performed, and tissue samples were harvested for histologic analyses. The gross and vascular anatomy of the pigs' spinal nerve roots were studied by dissection and ink injections. RESULTS There was a statistically significant decrease in the nerve conduction velocity in compressed compared with noncompressed spinal nerve roots after 1 week and after 4 weeks. The ameroid constrictors induced nerve fiber damage, endoneural hyperemia, bleeding, and inflammation at the compression zone. There was often a severe reduction in the number of myelinated fibers after 4 weeks. CONCLUSION A model for controlled, chronic, partial nerve root injury using a gradual compression-onset constrictor is presented. This model allows for induction of a controlled graded chronic nerve root injury and can be used for research on basic pathophysiologic mechanisms and on the effects of various interventions on nerve root injury development.

Study design Autologous nucleus pulposus was modified and applied to the cauda equina in pigs. Histology and neurophysiology were assessed after 7 days. Objectives To assess if alterations of the nucleus pulposus would change the degree and distribution of the nerve injury induced by autologous nucleus pulposus. Summary of background data It was reported recently that nucleus pulposus may induce structural and functional changes in nerve roots after epidural application. The basic mechanisms causing these changes are not fully understood. Methods Nucleus pulposus was harvested from lumbar discs and submitted to either of three treatments; 37 C for 24 hours (n = 5), -20 C for 24 hours (n = 5), or digestion by hyaluronidase for 24 hours (n = 6). In two additional pigs, nucleus pulposus was applied just after harvest as a control to verify previous observations. After 7 days, nerve conduction velocity was recorded, and specimens were processed for blinded light microscopic assessment. Results When nucleus pulposus was applied just after harvest, or when it had been kept at 37 C or digested by hyaluronidase for 24 hours, there was a significant reduction in nerve conduction velocity similar to previous observations. When nucleus pulposus had been kept at -20 C for 24 hours, however, there was no reduction in conduction velocity. There were no apparent differences between the groups at the histologic assessment. Staining of the nucleus pulposus showed that the cells in the nucleus pulposus exposed to -20 C were lysed, whereas the cells in the nucleus pulposus treated by the two other methods were mainly unaffected. Conclusions Because freezing of the nucleus pulposus probably kills the cells but does not affect other components, one may assume that the biologic effects induced by the nucleus pulposus may be related to its cell population.

Study design Intramuscular pressure and blood flow of the back muscles were evaluated topographically during posterior lumbar spine surgery. The topographic damage of the back muscle after surgery was studied. Objective To investigate the relationship between intramuscular pressure or blood flow during posterior lumbar surgery and the back muscle injury after surgery. Summary of background data Latrogenic back muscle injury in an animal and human model has been reported previously. Changes of intramuscular pressure and blood flow during surgery might be related to the muscle injury. No previous study on this issue has been published. Methods The contact pressure between the retractor blade and muscle tissue was monitored in 10 pigs during posterior surgery of the lumbar spine. On one side, intramuscular pressure at 5, 10, and 20 mm lateral to the retractor and on the other side blood flow of the back muscle at 5 and 20 mm during surgery were measured. Histologic changes of the back muscle at 5, 10, and 20 mm to the midline were evaluated 3 hours after surgery. Results The contact pressure decreased exponentially with time. Intramuscular pressure 5 mm lateral to the retractor was 114 +/- 31 mm Hg and was significantly higher than at 10 mm and 20 mm. Blood flow markedly decreased during surgery and recovered incompletely after releasing the retractor at 5 mm and 20 mm lateral to the retractor. Blood flow at 5 mm was significantly lower than at 20 mm throughout surgery. The muscle damage 3 hours after surgery was more severe near the retractor blade. Conclusions The back muscles were exposed to pathophysiologic condition by a retractor during posterior lumbar spine surgery. External compression by a retractor increases intramuscular pressure to levels that impede local muscle blood flow. The muscle degeneration after surgery could be explained by direct mechanical damage and by the increased intramuscular pressure of muscle tissue by the retractor.

Study Design. Lumbar intertransverse process arthrodesis using recombinant human bone morphogenetic protein-2 was performed in a previously established rabbit model for posterolateral spinal fusion and compared with fusions achieved using autogenous bone graft. Objectives. To qualitatively compare different recombinant human bone morphogenetic protein-2 dosages and carriers and to determine the efficacy of recombinant human bone morphogenetic protein-2 as a bone graft substitute to produce lumbar intertransverse process fusion in a validated rabbit model for posterolateral spinal fusion. of Background Data. Autogenous bone was considered the most successful bone graft material used for spinal arthrodesis. Problems with its use may occur in 25-30% of patients and prompted the search for and investigation of bone graft substitutes and osteoinductive growth factors, such as bone morphogenetic proteins. Recombinant human bone morphogenetic protein-2 was used successfully in orthotopic sites to generate bone in animal mandibular and long bone defect models. Methods. Posterolateral intertransverse process arthrodeses were performed at L5-L6 in 56 rabbits using recombinant human bone morphogenetic protein-2 or autogenous bone graft. Rabbits were killed either 5 weeks later to qualitatively compare fusions achieved using different recombinant human bone morphogenetic protein-2 dosages and carriers or 4 weeks later to compare the efficacy of recombinant human bone morphogenetic protein-2 in achieving spinal fusion compared with using autogenous bone graft. Inspection, manual palpation, radiography, histology, and biomechanic testing were used to assess the fusion. Results. All rabbits implanted with recombinant human bone morphogenetic protein-2 achieved solid spinal fusion by manual palpation and were fused radiographically, whereas only 42% of the autograft control fusions were solid. More mature fusions with greater trabecular bone formation were shown radiographically and histologically in rabbits implanted with the high-dose recombinant human bone morphogenetic protein-2 than with the low-dose recombinant human bone morphogenetic protein-2. Fusions achieved using recombinant human bone morphogenetic protein-2 delivered in the collagen carrier were more remodeled and homogeneous compared with using recombinant human bone morphogenetic protein-2 delivered in autograft ± collagen carrier. Fusions achieved with recombinant human bone morphogenetic protein-2 were biomechanically stronger and stiffer than fusions achieved using autogenous bone graft. Conclusions. Recombinant human bone morphogenetic protein-2 successfully and reliably achieved lumbar intertransverse process fusion in a validated rabbit model for posterolateral spinal fusion. Radiographically and histologically, greater and more rapid bone formation, consolidation, and remodeling were shown with recombinant human bone morphogenetic protein-2 compared with autogenous bone graft. Fusions achieved with recombinant human bone morphogenetic protein-2 were biomechanically stronger and stiffer than autograft fusions.

Study Design. The inflammatogenic properties of nucleus pulposus were assessed in two experimental models previously used for screening of inflammatogenic properties of other substances. This study was performed to assess the inflammatogenic properties of nucleus pulposus in models previously screened for other substances. Summary of Background Data. Previous experimental studies, as well as clinical observations, have indicated that inflammatory mechanisms may constitute an important pathogenetic component in sciatica due to herniation of the nucleus pulposus. Methods. In the first experimental series, autologous nucleus pulposus and retroperitoneal fat were placed in perforated titanium chambers, which were placed subcutaneously in pigs, together with an empty chamber as sham. After 7 days, the number of leukocytes in the chambers was determined. In the second experimental series, the microvascular reactions were studied by vital microscopy of the hamster cheek-pouch after local injections of suspensions of homologous nucleus pulposus and homologous subcutaneous fat. Macromolecular extravascular leakage was studied by fluorescence microscopy using FITC-dextran as a tracer. Results. The leukocyte ratio betwee fat control and sham was 0.9±0.6 and between nucleus pulposus and sham 2.4±0.7. The nucleus pulposus thus attracted significantly more leukocytes than fat. Injection of nucleus pulposus suspension induced thrombosis formation and pronounced leakages of macromolecules in a majority of the injection sites. However, injection of vehicle and fat suspension in the cheek-pouch only resulted in minor vital microscopic changes. Conclusions. Nucleus pulposus demonstrated inflammatogenic properties as indicated by leukotaxis and an increase of vascular permeability. It was not clear, however, if these reactions were induced by substances from the nucleus pulposus per se or from substances being liberated from other as a response to an interaction with components of the nucleus pulposus

Study design The possibility of epidurally applied substances reaching the intraneural capillaries of the spinal nerve roots and cauda equina was assessed in the pig sacrococcygeal spine. Methods The presence of Evans blue-labelled albumin in intraneural capillaries after epidural application for 1, 10, or 30 minutes was studied with fluorescence microscopy. Ink angiography was used to determine whether there were any direct communicating vessels between the epidural vein plexus and the intraneural capillaries. Results Evans blue-labelled albumin was present in the intraneural capillaries 1 minute after epidural application. Microangiography demonstrated small venules that connected the epidural vein plexus and the intraneural capillaries. Conclusions The results of this study demonstrated a rapid transport route between the epidural space and the intraneural capillaries. The results suggest that nucleus pulposus material, as well as epidurally applied substances, such as local anesthetic drugs or epidurally injected corticosteroids, may have a rapid, direct transport route to the axons of the spinal nerve roots. The demonstrated transport route also may be related to the mechanisms behind epidural anesthesia and spinal nerve root infiltration.

Study Design. The effects of intervention by intravenous injection of methylprednisolone to reduce the nerve root injury after epidural application of autologous nucleus pulposus was studied in an experimental model on the pig cauda equina in 20 animals. Methods. Nucleus pulposus was harvested from a lumbar disc. After lowering the pH of the nucleus pulposus to 3.5 it was placed onto the sacrococcygeal cauda equina. Fifteen of the pigs received a single intravenous injection of 30 mg/kg methylprednisolone, 5 minutes, 24 hours, or 48 hours, respectively, after the application. After 7 days, the nerve conduction velocity was determined, and biopsies of the cauda equina was examined by lightmicroscopy. Results. In the live pigs that did not receive any methylprednisolone treatment, nerve conduction velocity was reduced, whereas it was normal in the pigs treated 5 minutes and 24 hours after nucleus pulposus application. In pigs treated after 48 hours, nerve conduction velocity was reduced only slightly. At the light microscopic level, significant changes occurred in all series. Conclusions. This study indicates that the nucleus pulposus-induced effects on nerve function in an experimental pig model may be reduced dramatically by high-dose methylprednisolone administration within 24–48 hours after epidural application of autologous nucleus pulposus. The light microscopic changes were probably not significant for the nerve function. Instead, a morphologic explanation on a subcellular level should probably be sought.

Epidural application of autologous nucleus pulposus in pigs, without mechanical nerve root compression, induced a pronounced reduction in nerve conduction velocity in the cauda equina nerve roots after 1-7 days, compared to epidural application of retroperitoneal fat in control experiments. Histologically, the nerve fiber injury was more pronounced after application of nucleus pulposus than after control tissue application. The results demonstrate that nucleus pulposus may induce nerve tissue injury by mechanisms other than mechanical compression. Such mechanisms may be based on direct biochemical effects of nucleus pulposus components on nerve fiber structure and function and microvascular changes including inflammatory reactions in the nerve roots.

Double level compression of nerve roots has been known to exist for a long time, but only in recent years the clinical significance of such conditions has been discussed. In the current study, nerve root function in double level compression was analyzed by measurements of tail muscle electromyogram, after stimulations of 1) the intermediate segment between the compression sites and 2) cranial to the two compression sites in a pig model. The electromyogram-amplitudes were more reduced after intermediate, than cranial stimulation. This difference was suggested to be based on an additional increasing threshold to elicit an action potential within the intermediate segment due to a nutritional deficit at this location. The results further stress the importance of local changes within the intermediate segment in double level nerve root compression.

Double-level nerve root compression had a greater effect on nerve impulse conduction, recorded as tail muscle action potential amplitude, than single-level compression at both 10 and 50 mm Hg. If the distance between the two compression sites at 10 mm Hg compression was increased from one vertebral segment (10 mm) to two vertebral segments (30 mm), the reduction of muscle action potential amplitude was further enhanced. These observations may give new insight into the pathophysiology of cauda equina compression in spinal disorders such as stenosis.

Previous recommendations regarding the "safe" period of tourniquet hemostasis were based largely on studies of ischemia distal to the tourniquet. This study quantitatively analyzed skeletal muscle injury induced beneath and distal to a pneumatic tourniquet applied to the hindlimbs of rabbits for 1, 2, or 4 hours with a cuff inflation pressure of 125, 200, or 350 mm Hg. Technetium Tc 99m pyrophosphate incorporation after systemic injection (Tc 99 uptake) and correlative histology were used to evaluate tissue damage 2 days after tourniquet application. Compared with the contralateral control limbs, compression and ischemia induced statistically significant increases in Tc 99 uptake in the thigh and leg regions of all groups. Pyrophosphate incorporation was significantly greater in the thigh region than in the leg region after 2 hours of compression in the 200 and 350 mm Hg pressure groups and following 4 hours of compression in all pressure groups. Focal and regional fiber necrosis and degeneration were observed in thigh muscles after 2 hours of tourniquet compression. Two hours of continuous tourniquet application at clinically relevant cuff inflation pressures induced significant skeletal muscle necrosis beneath the tourniquet. Use of the lowest possible inflation pressure for a limited duration should minimize the degree of tissue injury caused by tourniquet application.

Neurophysiologic and neuropathologic changes were studied in rabbit hindlimbs after 2 hours of pneumatic tourniquet application with either 350 mmHg (n = 18) or 1,000 mmHg (n = 6) cuff inflation pressure. The toe spread reflex was decreased in 66% and absent in 33% of limbs 2 days after 350 mmHg compression, and was absent in all limbs after 1,000 mmHg compression. Compound motor action potential amplitudes (CMAPs), recorded from the abductor hallucis muscle, were significantly decreased with sciatic nerve stimulation 1 hour after 350 mmHg compression. CMAPs returned to baseline values one and two days later, however nerve conduction velocity (NCV) was still significantly decreased in the compressed sciatic nerves of these groups. In contrast, complete nerve conduction block, localized beneath the cuff's distal border, was observed two days after 1,000 mmHg compression, and NCV was still significantly decreased distal to the tourniquet zone. Using light and electron microscopy, scattered axonal degeneration, mild myelin damage, and normal nodes of Ranvier were observed two days after 350 mmHg tourniquet compression. Severe fiber damage and nodal obliteration were noted after 1,000 mmHg tourniquet compression. Although nodal invagination is probably not a significant pathogenic mechanism at clinically relevant tourniquet pressures and durations, functional abnormalities were induced by 2 hour, 350 mmHg tourniquet compression. Such changes probably correlate with clinical electromyographic abnormalities and delayed post-operative recovery following 'routine' extremity surgery using pneumatic tourniquets.

A well-controlled animal model is presented for the study of neuromuscular injury induced by a pneumatic tourniquet. This model comprises a curved tourniquet surrounded by a stiff exterior shell, both of which were specifically designed to fit the conical and oblong shape of the rabbit hindlimb. Computed tomographic imaging was used to assess transverse tissue displacement induced by tourniquet compression. The curved tourniquet/shell configuration occluded the distal arterial blood flow to the extremity at a significantly lower cuff inflation pressure than a straight tourniquet of equal width. The magnitude and distribution of tissue pressures in the subcutaneous and deep tissues beneath the tourniquet were similar to those recorded in previous human cadaver studies of tourniquet compression. This animal model will facilitate the quantitation and analysis of tissue injury induced beneath and distal to a pneumatic tourniquet. Such data can help define the critical pressure and time limits for the safe use of pneumatic tourniquets in extremity surgery.

Peripheral nerves are often mechanically stretched in association with trauma to extremities, resulting in varying degrees of impairment of nerve function. However, little is known about the biomechanical properties of peripheral nerves and limits of stretching that the nerve may undergo before structural changes occur. Also, the injury pattern of nerves under stretching is poorly understood. In the present study, fresh rabbit tibial nerves (n = 18) were harvested. Nine nerves were stretched to failure in an INSTRON materials testing machine at a rate of 1 cm/min (strain rate of 0.5%/s). Load-deformation and stress-strain curves were determined. Histological examination by light microscopy of the stretched nerves as well as six normal control nerves and three clamped nonstretched control nerves was performed. The results show that the rabbit tibial nerves have an in situ strain of 11.0 +/- 1.5% and exhibit a nonlinear stress-strain relationship. After 20% strain, the curve becomes linear up to failure. The ultimate strain and tensile strength of the nerves were 38.5 +/- 2.0% and 11.7 +/- 0.7 MPa, respectively. At failure, the load dropped suddenly, but the specimens remained grossly intact. Histological analysis of the stretched nerves showed multiple ruptures of perineurial sheaths when compared to controls.

The effects of compression on the transport of 3H-labeled methyl glucose to spinal nerve roots were analyzed in an experimental model of the pig cauda equina. A rapid onset of compression (0.05-0.1 s) induced more pronounced effects than a slow onset (20 s) at corresponding pressure levels. There was evidence that this observed difference may be related to the magnitude of intraneural edema formed outside the compression zone. The results also indicate that the nutritional transport might be impaired at very low pressure levels and that diffusion from adjacent tissues with a better nutritional supply, including the cerebrospinal fluid, may not fully compensate for any compression-induced impairment of the intraneural blood flow.

We describe the morphologic changes that follow division of the transverse carpal ligament in patients with carpal tunnel syndrome. Fifteen hands in 12 patients with carpal tunnel syndrome were studied with magnetic resonance imaging before operation and for 6 weeks after operation. Eight hands were studied at 8 months after operation. Carpal arch width, anterior displacement of the carpal canal contents, and carpal canal volume were measured by use of multiplanar reformation and three-dimensional reconstruction of magnetic resonance images. There was a 24.2 +/- 11.6% increase in carpal canal volume 6 weeks after carpal tunnel release (p less than 0.001). This difference persisted at 8-month follow-up. There was an anterior displacement of carpal canal contents 3.5 +/- 1.9 mm from its original position 6 weeks after operation (p less than 0.001). This palmar displacement persisted at the 8-month follow-up. There was no statistically significant increase in carpal arch width 8 months after carpal tunnel release. We believe that division of the transverse carpal ligament restores median nerve function by increasing the volume of the carpal canal. This volumetric increase results from an anterior displacement of the newly formed transverse carpal ligament and not from a widening of the bony carpal arch.

Compression applied to a peripheral nerve may easily interfere with intraneural blood flow. In the present experimental study, a vital microscopic technique was used to observe changes in intraneural microcirculation (intrafascicularly and extrafascicularly) when graded compression was applied to a rabbit's tibial nerve by a specially designed minicompression device. Interference with venular flow was observed already at a pressure of 20 to 30 mm Hg while arteriolar and intrafascicular capillary flow was impaired at about 40 to 50 mm Hg. At 60 to 80 mm Hg no blood flow could be observed in the nerve. Nerves observed 3 or 7 days after 2 hours of compression at 400 mm Hg showed no or very slow stagnant blood flow within the previously compressed segment. It is concluded that acute compression of nerve may cause persistent impairment of intraneural microcirculation due to mechanical injury to blood vessels.

Flexor tendon from the synovial sheath region of rabbits was resected and placed as a free body in a synovial knee joint recess. Structural changes of the surface of the graft caused by the instruments used to handle the tendon during the surgery (jewellers microforceps), was analysed by electron microscopy at various time intervals up to 6 weeks. The superficial coat of amorphous material, i.e. ground substance, on the surface of the tendon was dishevelled by the instruments. The injured areas were crowded with monocytes and macrophages during the first days. An increasing number of fibroblasts and mesothelial-like cells covered the damaged area in less than two weeks. The superficial coat of ground substance was not completely restored until after 3 weeks or more. The results obtained are discussed in relation to functional aspects.

The anatomy and physiology of the nerve root complex in the lumbar spine are reviewed, with special reference to the effects of mechanical deformation of nerve roots in association with intervertebral disc herniation and spinal stenosis. Biomechanical aspects of nerve root deformation induced by compression are discussed. The functional changes induced by compression can be caused by mechanical nerve fiber deformation but also may be a consequence of changes in nerve root microcirculation, leading to ischemia and formation of intraneural edema. Nerve root compression can, by different neurophysiologic mechanisms, induce motor weakness and altered sensibility or pain. Intraneural edema and demyelination seem to be critical factors for the production of pain in association with nerve root compression.

The minimal pressure for impairment of fast anterograde axonal transport was determined in rabbit vagus nerve. Proteins, transported by fast anterograde axonal transport, were labeled by a microinjection of [3H]leucine into the nodose ganglion, and a small compression chamber was applied around the cervical vagus nerve. In this way the nerve was subjected to acute, graded compression. Compression at 20 mm Hg for 2 h as well as sham compression did not induce accumulation of axonally transported proteins at the level of compression. However, a pressure of 30 mm Hg for 2 h induced a block of axonal transport at the site of compression. The causes of the axonal transport block are discussed as well as the minimal pressure level in relation to pressures found in clinical nerve compression lesions.

Severed and subsequently sutured rabbit flexor tendons were kept free and isolated in the synovial cavity of the knee joint. In one series the tendon specimens were surrounded by a dialyzing membrane to avoid cell seeding from the synovial fluid. At different intervals of time over a period of 3 weeks, the tendons were studied morphologically with special reference to scanning electron microscopy. Adhesions were not observed and, with synovial fluid as the nutrient medium, the tendons showed an intrinsic ability to repair in the superficial layers, also bridging the suture gap. Moreover, cell seeding, mainly of macrophages, from the synovial fluid could be demonstrated on the very surface of the tendon. When this cell seeding was prevented, the fibroplasia in the superficial layer of the tendon did decrease slightly, but tendon cell morphology was that of active fibroblasts. The results support the concept that flexor tendons may show intrinsic fibroplasia when nourished by synovial fluid, while macrophages, mainly, of extrinsic origin, contribute to restoration of the tendon surface.