Using a regional arthroplasty register we assessed the outcome at five years of 1198 primary Charnley total hip replacements (THRs) carried out in 1152 patients across a single UK health region in 1990. Information regarding outcome was available for 1080 hips (90%) and 499 had an independent clinical and radiological assessment.

By five years the known rate of aseptic loosening was 2.3%, of deep infection 1.4%, of dislocation 5.0% and of revision 3.2%. The radiological assessment of 499 THRs revealed gross failure in a further 5.2%, which had been previously unrecognised. The combined rate of failure of nearly 9% is higher than those published from specialist centres and surgeons, but is probably more representative of the norm.

Our study supports the need for a national register and surveillance of THRs. It emphasises that all implants should be followed, and suggests that the results of such surgery, when performed in the general setting, may not be as good as expected.

We have followed up for a period of seven to nine years 100 consecutive arthroplasties of the hip in which an entirely HA-coated implant had been used.

The clinical results were excellent and bony incorporation was extensive in all components. No stem became loose or subsided but five cups were revised because of loosening after 3.8 to 5.5 years, having functioned painlessly and shown radiological ingrowth.

Revision procedures because of excessive polyethylene wear have been performed on 18 hips and are planned for six more. Two eroded metal backings with worn-through polyethylene were exchanged; six hips showed metallosis without polyethylene wear-through. There were two cases of granulomatous cysts in the groin and 66 hips had osteolysis located periarticularly, in the greater trochanter or in the acetabulum.

Six major and seven minor diagnostic criteria have been developed by the Japanese Investigation Committee for osteonecrosis of the femoral head (ONFH). We have carried out a multicentre study to clarify these.

We studied prospectively 277 hips in 222 patients, from six hospitals, who had ONFH and other hip pathology and from whom histological material was available. We identified five criteria with high specificity: 1) collapse of the femoral head without narrowing of the joint space or acetabular abnormality on radiographs, including the crescent sign; 2) demarcating sclerosis in the femoral head without narrowing or acetabular abnormality; 3) a ‘cold-in-hot’ appearance on the bone scan; 4) a low-intensity band on T1-weighted images (band pattern); and 5) evidence of trabecular and marrow necrosis on histological examination. With any combination of two of these criteria, the sensitivity and specificity of the diagnosis were 91% and 99%, respectively.

We performed dynamic MRI of the femoral head within 48 hours of injury on 22 patients with subcapital fracture of the neck of the femur and on a control group of 20 of whom ten were healthy subjects and ten were patients with an intertrochanteric fracture. Three MRI patterns emerged when the results between the fractured side and the contralateral femoral head were compared. In all of the control group and in those patients who had undisplaced fractures (Garden stages I and II), perfusion of the femoral head was considered to be at the same level as on the unaffected side. In patients with displaced fractures (Garden stages III and IV) almost all the femoral heads on the fractured side were impaired or totally avascular, although some had the same level of perfusion as the unaffected side. We conclude that dynamic MRI, a new non-invasive imaging technique, is useful for evaluating the perfusion of the femoral head.

We describe a prospective study of 328 patients undergoing arthroscopy of the hip. Arthrography, CT or MRI was performed when clinically indicated. A preoperative diagnosis was reached in 174 patients (53%), while the remaining 154 were diagnosed as having ‘idiopathic hip pain’.

In seven patients, access to the hip was inadequate. Arthroscopy altered the diagnosis in 176 hips (53%). The new primary diagnoses were osteoarthritis (75 patients), osteochondral defects (34), torn labra (23), synovitis (11) and loose bodies (9). In 172 hips (52%) an operative procedure was undertaken. In the remaining 84 patients (26%), arthroscopy neither changed the diagnosis nor provided surgical treatment.

Arthroscopy of the hip is considered to be of value in assessing and treating the adult patient with pain in the hip of uncertain cause.

Although arthroscopy of the hip is being carried out increasingly, little is known about the rate of associated complications. We describe a prospective study of 640 consecutive procedures in which a consistent technique was used. The overall complication rate was 1.6%. Complications, none of which was major or long-term, included transient palsy of the sciatic and femoral nerves, perineal injury, bleeding from the portal wounds, trochanteric bursitis and intra-articular breakage of the instrument. We believe that it is possible to undertake this operation safely using the technique described.

We describe in 30 feet the occurrence of a tarsal tunnel syndrome caused by a ganglion. The presenting symptom was numbness or pain in the toes and the sole with paraesthesiae in the distribution of the medial plantar nerve in 63% of the patients. Swellings which were not palpable were detected by ultrasonography. Twenty-nine patients were treated by operation. Most ganglia originated from the talocalcaneal joint, and five were associated with a talocalcaneal coalition. The surgical outcome was satisfactory in all patients except one who had a further operation for a recurrence of the ganglion.

We have reviewed the long-term results of 22 patients (23 fusions) with fractures of the os calcis, who had subtalar arthrodesis with correction of the deformity between 1975 and 1991. The mean follow-up was nine years (5 to 20).

All patients were evaluated according to a modified foot score. A radiological assessment was used in which linear and angular variables were measured including the fibulocalcaneal abutment, the height of the heel and fat pad, the angle of the arch and the lateral talocalcaneal and the lateral talar declination angles.

The technique used restores the normal relationship between the hindfoot and midfoot and corrects the height of the heel. This leads to better biomechanical balance of the neighbouring joints and gives a favourable clinical outcome.

The modified foot score showed a good or excellent result in 51% of the feet. Residual complaints were mostly due to problems with the soft tissues. Subjectively, an excellent or good score was achieved in 78% of the cases.

After statistical analysis, except for the height of the heel and the degenerative changes in the calcaneocuboid joint, no significant difference was found in the measured variables between the operated and the contralateral side.

Disarticulation has been carried out in ten ankles in nine patients in whom it was not possible to use a heel flap. Four patients were able to walk with a prosthesis which gave satisfactory function. In five who were bedridden, healing was achieved and was of sufficient quality to allow transfers. There was no operative morbidity or mortality.

This technique can be used instead of a transtibial amputation if necrosis or ischaemia of the heel is a contraindication to conventional Syme’s amputation.

We studied prospectively 30 patients who had a Mitchell’s osteotomy secured by either a suture followed by immobilisation in a plaster boot for six weeks, or by a cortical screw with early mobilisation.

The mean time for return to social activities after fixation by a screw was 2.9 weeks and to work 4.9 weeks, which was significantly earlier than those who had stabilisation by a suture (5.7 and 8.7 weeks, respectively; p < 0.001). Use of a screw also produced a higher degree of patient satisfaction at six weeks, and an earlier return to wearing normal footwear. The improvement in forefoot scores was significantly greater after fixation by a screw at six weeks (p = 0.036) and three months (p = 0.024). At one year, two screws had been removed because of pain at the site of the screw head.

Internal fixation of Mitchell’s osteotomy by a screw allows the safe early mobilisation of patients and reduces the time required for convalescence.

We examined clinically and radiologically the knees of 46 patients (27 females and 19 males) with diastrophic dysplasia. The age of the patients varied from newborn to 38 years. A total of 18 patients was followed during their growth until adolescence. The knees of two legally aborted fetuses appeared on examination to be macroscopically normal and congruous. Excessive valgus deformity of the tibiofemoral weight-bearing angle with a mean of 14° was noted in infancy. Most of the patients had marked instability of the knees. The range of movement of the knee began to decrease before the age of five years. There were signs of early degeneration and deformation of the bony epiphyses before the age of six years. The patellofemoral joint was abnormal from an early age. A marked patella infera, often associated with a lateral position of the patella with bony fragmentation, was noted. The knee in diastrophic dysplasia is basically unstable, showing early deformation of the subchondral bone and degeneration of the joint.

We took posteroanterior weight-bearing radiographs, both with the joint fully extended and in 30° of flexion, in a consecutive series of 50 knees in 37 patients referred for the primary assessment of pain and/or stiffness. These radiographs were reported ‘blind’ both by an orthopaedic surgeon and a radiologist. Direct measurement of the joint space, together with grading of the severity of erosion according to the Ahlback criteria, was undertaken. Any other abnormality present was also documented.

The radiographs of the knees in 30° of flexion consistently showed more advanced erosion in both the medial (p = 0.001) and the lateral (p = 0.0001) tibiofemoral compartments, when compared with those of knees in full extension. The Ahlback classification of 25 joints was altered, in some cases by several grades, by the flexed position of the joint. In every case in which another abnormality was identified on the radiograph in full extension, it was also noted on that of the knee in 30° of flexion. In a further four cases, additional pathology could only be seen in the flexed knee. Every patient was able to complete the radiological examination without difficulty. Our study supports the adoption of a weight-bearing view in 30° of flexion as the standard posteroanterior radiograph for the assessment of tibiofemoral osteoarthritis in patients over 50 years of age.

We designed an experimental study to prove the existence of the popliteofibular ligament (PFL) and to define its role in providing static stability of the knee. We also examined the contribution of the lateral collateral ligament (LCL).

We found this ligament to be present in all eight human cadaver knees examined. These specimens were mounted on a specially designed rig and subjected to posterior, varus and external rotational forces. We used the technique of selective sectioning of ligaments and measured the displacement with a constant force applied, before and after its division. We recorded the displacement in primary posterior translation, coupled external rotation, primary varus angulation and primary external rotation. Statistical analysis using the standard error of the mean by plotting 95% confidence intervals, was used to evaluate the results.

The PFL had a significant role in preventing excessive posterior translation and varus angulation, and in restricting excessive primary and coupled external rotation. Isolated section of the belly of popliteus did not cause significant posterolateral instability of the knee. The LCL was also seen to act as a primary restraint against varus angulation and secondary restraint against external rotation and posterior displacement. Our findings showed that in knees with isolated disruption of the PFL stability was restored when it was reconstructed. However in knees in which the LCL was also disrupted, isolated reconstruction of the PFL did not restore stability.

We report a patient with a subperiosteal ganglion cyst of the tibia which was imaged by radiography, arthrography, CT and MRI. The images were correlated with the arthroscopic surgical and histological findings. Spiculated formation of periosteal new bone on plain radiographs led to the initial suspicion of a malignant tumour. Demonstration of the cystic nature of the tumour using cross-sectional imaging was important for the precise diagnosis. Communication between the ganglion cyst and the knee was shown by a delayed arthrographic technique, and the presence of this communication was confirmed at arthroscopy and surgically.

We reviewed the outcome of 146 Insall-Burstein II total knee replacements carried out in 121 patients over a period of nearly four years in a general orthopaedic unit. At a mean follow-up of ten years, 94 knees in 78 patients were available for review. Six patients (7 knees) were lost to follow-up and 37 (45 knees) had died.

The clinical outcome using the scoring system of the Hospital for Special Surgery (HSS) was excellent or good in 79% of patients, fair in 14% and poor in 7%. The mean preoperative HSS score was 31, improving to 79 at the latest review. Using the newer rating system of the Knee Society, the mean score at ten years was 87 and the mean functional score 56. The arc of flexion improved from a mean preoperative value of 88° to 100°. The 18 patients who had had a previous high tibial osteotomy were analysed separately and were found to have benefited equally from the operation. Nine prostheses were revised, giving a cumulative survival rate of 92.3% at ten years. Radiological evaluation of 104 radiographs showed radiolucent lines around ten tibial components, none of which required revision. Anterior knee pain was a significant problem.

The optimal regime of antithrombotic prophylaxis for patients undergoing total knee arthroplasty (TKA) has not been established. Many surgeons employ intermittent pneumatic compression while others use low-molecular-weight heparins (LMWH) which were primarily developed for total hip arthroplasty. We compared the efficacy and safety of these two techniques in a randomised study with blinded assessment of the endpoint by phlebography.

We randomised 130 patients, scheduled for elective TKA, to receive one daily subcutaneous injection of nadroparin calcium (dosage adapted to body-weight) or continuous intermittent pneumatic compression of the foot by means of the arteriovenous impulse system.

A total of 108 patients (60 in the LMWH group and 48 in the mechanical prophylaxis group) had phlebography eight to 12 days after surgery. Of the 47 with deep-vein thrombosis, 16 had received LMWH (26.7%, 95% CI 16.1 to 39.7) and 31, mechanical prophylaxis (64.6%, 95% CI 49.5 to 77.8). The difference between the two groups was highly significant (p < 0.001). Only one patient in the LMWH group had severe bleeding.

We conclude that one daily subcutaneous injection of calcium nadroparin in a fixed, weight-adjusted dosage scheme is superior to intermittent pneumatic compression of the foot for thromboprophylaxis after TKA. The LMWH scheme was also safe.

Between 1992 and 1994 we performed a prospective study of the effect of total knee replacement (TKR) on the health status of 119 patients over the age of 80 years who had had a primary unilateral TKR.

The Nottingham Health Profile was used to assess this before and at three and 12 months after operation. We found a significant improvement in the scores for pain, emotional reaction, sleep and physical mobility at three months. After 12 months, the scores for pain and sleep were well maintained. The other factors had deteriorated slightly but remained better than before operation.

Our findings show that TKR leads to a significant improvement in the general health status of the very elderly.

In two years we treated four women with ununited stress fractures of their proximal tibial diaphyses. They all had arthritis and valgus deformity. The stress fractures had been treated elsewhere by non-operative means in three patients and by open reduction and internal fixation in one, but had failed to unite. After treatment with a modular total knee prosthesis with a long tibial stem extension, all the fractures united.

A modular total knee prosthesis is suitable for the rare and difficult problem of ununited tibial stress fractures in patients with deformed arthritic knees since it corrects the deformity and the adverse biomechanics at the fracture site, stabilises the fracture and treats the arthritis.

We report two cases of vertebral osteochondroma. In one patient a solitary cervical lesion presented as entrapment neuropathy of the ulnar nerve and in the other as a thoracic tumour associated with hereditary multiple exostoses producing paraplegia. We highlight the importance of an adequate preoperative evaluation in such patients.

Degenerative spondylolisthesis is four times more common in women than in men. Although this gender difference has long been recognised there has been no explanation for it. We have examined the radiographs and CT scans of 118 patients over the age of 55 years and of a control group under the age of 46 years.

Our findings confirmed the presence of more sagittally-orientated facet joints in patients with degenerative spondylolisthesis but did not show that the gender difference can be explained by the morphology of the facet joint.

Furthermore, we conclude that the increased angle of the facet joint is the result of arthritic remodelling and not the primary cause of degenerative spondylolisthesis. It is more likely to be due to loss of soft-tissue resilience with subsequent failure of the facet joints which are acting as the last restraints to subluxation.

The diagnosis of chronic lesions of the rotator cuff is challenging. We have developed a new index to improve the sonographic diagnosis of chronic tears of the cuff.

In a pilot study, we examined 50 asymptomatic healthy volunteers by ultrasound to establish the diameter of the rotator cuff in relation to the tendon of the long head of biceps. Subsequently, the index was calculated in 64 patients who had had shoulder pain for more than three months caused by clinically diagnosed lesions of the rotator cuff. The compensatory hypertrophy of the biceps tendon was quantified sonographically in relation to the diameter of the cuff.

Comparison with the contralateral shoulder revealed a significantly higher biceps rotator-cuff ratio (p < 0.05) for patients with torn rotator cuffs. A ratio greater than 0.8 was considered pathological (index positive); the mean ratio in the control group was 0.43. The sensitivity of a positive index was 97.8%, the specificity 63.2%, the positive predictive value 86.3%, and the negative predictive value 92.4% in comparison with surgical findings.

Use of the index improves sensitivity in the diagnosis of chronic tears of the cuff by ultrasound.

Opinion varies as to the incidence of nerve lesions in anterior dislocation of the shoulder after low-velocity trauma. Most studies are retrospective or do not use EMG. We have investigated the incidence and the clinical consequences of nerve lesions in a prospective study by clinical and electrophysiological examination.

Axonal loss was seen in 48% of 77 patients. The axillary nerve was most frequently involved (42%). Although recovery as judged by EMG and muscle strength was almost complete, function of the shoulder was significantly impaired in patients with lesions of the axillary and suprascapular nerves. Unfavourable prognostic factors are increasing age and the presence of a haematoma.

It is not necessary to carry out EMG routinely; an adequate programme of physiotherapy is important. In patients with a severe paresis, EMG is essential after three weeks.

Heritable thrombophilic disorders have been proposed as one of the causes for Legg-Calvé-Perthes disease. A total of 62 patients diagnosed with this disease between 1988 and 1997 and 50 controls were screened for thrombophilia. The incidence and relationship of thrombophilia to the severity of the disease were evaluated.

One patient and none of the controls had protein S deficiency. One of the control group and one of the patients had protein C deficiency with the latter child also having a combined deficiency with a mutant factor V gene.

The number of children with a mutant factor V gene, protein C deficiency, who were homozygous for the C 677T polymorphism of methylenetetra-hydrofolate reductase or were heterozygous for mutant G20210A prothrombin did not differ statistically in the study and the control groups. No patient had antithrombin deficiency or positive lupus anticoagulant.

We found no correlation between thrombophilia and the extent of the disease. The most common risk factors for arteriovenous thromboembolism showed no statistical significance in our patients compared with the control group or with the general population. These data do not confirm an aetiological role for thrombophilia in Perthes’ disease.

Recent work has suggested that thrombophilia may be an aetiological factor in up to 50% of children with Perthes’ disease, and that up to 75% may have a coagulopathy. Our aim was to test these findings in the local population of children with Perthes’ disease and attempt to correlate them with the severity of the condition.

In 64 children there were only eight (12%) with low levels of clotting proteins, as defined by normal paediatric ranges. Of these eight, only five could be said to show any thrombophilic tendency.

In order to define the prognostic factors in Perthes’ disease in children older than 12 years, we reviewed 15 patients at the end of growth who were aged 12.1 to 14 years at presentation. The patients with the worst long-term prognosis (Stulberg class V) were compared with the others for age, skeletal maturity and remaining growth (Oxford method), as well as Catterall and Waldenström classifications at presentation. A significant difference (p = 0.001) was found for remaining growth (25% in Stulberg class V and 35% in the others) and also for the results at the end of growth when the remaining growth was over 30%, since this allowed sufficient time for reformation and remodelling of the femoral head.

We have studied the relationship between movements of the foot and ankle and venous blood flow from the lower limb using colourflow Duplex ultrasound to determine the optimum type of exercise for promoting venous return. Studies of both active and passive movements were carried out on 40 limbs in 20 subjects (18 men; 2 women), with a median age of 27 years (20 to 54). We assessed ankle dorsiflexion and plantar flexion, subtalar inversion and eversion, and a combination of all movements.

There was no difference in venous flow when comparing opposite limbs in a single subject (p > 0.5), but active exercises produced higher peak and mean velocities of blood flow than passive ones. The active combined movement produced the highest velocities with an increase of 38% in mean and of 58% in peak flow velocities, which were significantly greater than the peak and mean flow rates produced by passive movements. The active combined exercise would therefore be the most effective in eliminating stasis and could contribute to the prevention of deep-vein thrombosis.

We have studied the effect of hydroxyapatite (HA) coating in 15 ovariectomised and 15 normal rats which had had a sham procedure. Twenty-four weeks after operation, HA-coated implants were inserted into the intramedullary canal of the right femur and uncoated implants into the left femur. The prostheses were removed four weeks after implantation. Twelve specimens in each group had mechanical push-out tests. Sagittal sections of the other three were evaluated by SEM.

The bone mineral density (BMD) of the dissected left tibia was measured by dual-energy x-ray absorptiometry. The difference in BMD between the control and ovariectomised tibiae was 35.01 mg/cm² (95% CI, 26.60 to 43.42). The push-out strength of the HA-coated implants was higher than that of the uncoated implants in both groups (p < 0.0001), but the HA-coated implants of the ovariectomised group had a reduction in push-out strength of 40.3% compared with the control group (p < 0.0001).

Our findings suggest that HA-coated implants may improve the fixation of a cementless total hip prosthesis but that the presence of osteoporosis may limit the magnitude of this benefit.

We performed a prospective, randomised double-blind study in 24 patients undergoing high tibial osteotomy to evaluate the effectiveness of human recombinant osteogenic protein (OP-1) on a collagen type-I carrier in a critically-sized fibular defect. The study had two phases, each evaluated by clinical, radiological and DEXA methods during the first postoperative year. The first concerned the validation of the model of the fibular defect, using positive (demineralised bone) and negative (untreated) controls. The second phase concerned the osteogenic potential of OP-1 on collagen type-I ν collagen type-I alone.

The results of the first phase established the critically-sized nature of the defect. In the untreated group no bony changes were observed while, in the demineralised bone group, formation of new bone was visible from six weeks onwards. The results of the second phase showed no significant formation of new bone in the presence of collagen alone, while in the OP-1 group, all patients except one showed formation of new bone from six weeks onwards. This proved the osteogenic activity of OP-1 in a validated critically-sized human defect.

Limited success in regenerating large bone defects has been achieved by bridging them with osteoconductive materials. These substitutes lack the osteogenic and osteoinductive properties of bone autograft. A direct approach would be to stimulate osteogenesis in these biomaterials by the addition of fresh bone-marrow cells (BMC).

We therefore created osteoperiosteal gaps 2 cm wide in the ulna of adult rabbits and either bridged them with coral alone (CC), coral supplemented with BMC, or left them empty. Coral was chosen as a scaffold because of its good biocompatibility and resorbability. In osteoperiosteal gaps bridged with coral only, the coral was invaded chiefly by fibrous tissue. It was insufficient to produce union after two months. In defects filled with coral and BMC an increase in osteogenesis was observed and the bone surface area was significantly higher compared with defects filled with coral alone. Bony union occurred in six out of six defects filled with coral and BMC after two months. An increase in the resorption of coral was also observed, suggesting that resorbing cells or their progenitors were present in bone marrow and survived the grafting procedure. Our findings have shown that supplementation of coral with BMC increased both the resorption of material and osteogenesis in defects of a clinical significance.

We inserted two hydroxyapatite (HA)-coated implants with crystallinities of either 50% (HA-50%) or 75% (HA-75%) bilaterally into the medial femoral condyles of the knees of 16 dogs. The implants were allocated to two groups with implantation periods of 16 and 32 weeks. They were weight-bearing and subjected to controlled micromovement of 250 μm during each gait cycle. After 16 weeks, mechanical fixation of the HA-50% implants was increased threefold as compared with the HA-75% implants. After 32 weeks there was no difference between HA-50% and HA-75%. Fixation of HA-75% increased from 16 to 32 weeks whereas that of HA-50% was unchanged. HA-50% implants had 100% more bone ingrowth than HA-75% implants after 16 weeks. More HA coating was removed on HA-50% implants compared with HA-75% implants after both 16 and 32 weeks. No further loss of the HA coating was shown from 16 to 32 weeks.

Our study suggests that the crystallinity of the HA coating is an important factor in its bioactivity and resorption during weight-bearing conditions. Our findings suggest two phases of coating resorption, an initial rapid loss, followed by a slow loss. Resorbed HA coating was partly replaced by bone ingrowth, suggesting that implant fixation will be durable.

Dupuytren’s disease is a chronic inflammatory process which produces contractures of the fingers. The nodules present in Dupuytren’s tissue contain inflammatory cells, mainly lymphocytes and macrophages. These express a common integrin known as VLA4. The corresponding binding ligands to VLA4 are vascular cell adhesion molecule-1 (VCAM-1) present on the endothelial cells and the CS1 sequence of the fibronectin present in the extracellular matrix. Transforming growth factor-beta (TGF-ß) is a peptide hormone which has a crucial role in the process of fibrosis.

We studied tissue from 20 patients with Dupuytren’s disease, four samples of normal palmar fascia from patients undergoing carpal tunnel decompression and tissue from ten patients who had received perinodular injections of depomedrone into the palm five days before operation. The distribution of VLA4, VCAM-1, CS1 fibronectin and TGF-ß was shown by immunohistochemistry using an alkaline phosphorylase method for light microscopy.

In untreated Dupuytren’s tissue CS1 fibronectin stained positively around the endothelial cells of blood vessels and also around the surrounding myofibroblasts, principally at the periphery of many of the active areas of the Dupuytren’s nodule. VCAM-1 stained very positively for the endothelial cells of blood vessels surrounding and penetrating the areas of high nodular activity. VCAM-1 was more rarely expressed outside the blood vessels. VLA4 was expressed by inflammatory cells principally in and around the blood vessels expressing VCAM-1 and CS1 but also on some cells spreading into the nodule. TGF-ß stained positively around the inflammatory cells principally at the perivascular periphery of nodules. These cells often showed VLA4 expression and co-localised with areas of strong production of CS1 fibronectin.

Normal palmar fascia contained only scanty amounts of CS1 fibronectin, almost no VCAM-1 and only an occasional cell staining positively for VLA4 or TGF-ß.

In the steroid-treated group, VCAM-1 expression was downregulated in the endothelium of perinodular blood vessels and only occasional inflammatory cell expression remained. Expression of CS1 fibronectin was also much reduced but still occurred in the blood vessels and around the myofibroblast stroma. VLA4-expressing cells were also reduced in numbers. A similar but reduced distribution of production of TGF-ß was also noted.

Our findings show that adherence of inflammatory cells to the endothelial wall and the extravasation into the periphery of the nodule may be affected by steroids, which reduce expression of VCAM-1 in vivo. This indicates that therapeutic intervention to prevent the recommencement of the chronic inflammatory process and subsequent fibrosis necessitating further surgery may be possible.