Search

Objective Primary malignant tumours of the bony chest wall are uncommon and data concerning treatment and results are sparse. Methods To assess the results of surgical resection and chest wall reconstruction we reviewed our experience with primary malignant chest wall tumours treated since 1958. Results Of the 49 lesions, 42 were found in the ribs and the remaining 7 in the sternum. These included chondrosarcomas [22], solitary plasmacytoma [18], Ewing's tumours [7], Askin's tumour [1] and Desmoid tumour [1]. Skeletal reconstruction was performed in 36 of the 49 patients. Marlex mesh alone was used in 17 patients. Since 1972, a sandwich of two layers of Marlex mesh with a filler of methyl methacrylate was utilised [19] successfully, producing better functional and cosmetic results. Primary soft tissue closure was possible in all but 8 cases in whom latissimus dorsi myocutaneous flaps were used. Bilaterally, partially transposed pectoralis major muscle was used to cover upper sternal defects in 4 cases. All but 1 patient had an uneventful post-operative recovery none requiring ventilatory support. Survival Overall survival at 5 and 10 years was 68%. The differential figures for 10-year survival were for chondrosarcoma 67%, Ewing's sarcoma 43%, and solitary plasmacytoma 59%. These were the results of radical en-bloc excisions. The patient with Desmoid tumour is alive at 5 years, following incomplete initial resection and the patient with Askin's tumour survived for 3 years. Conclusion Radical en-bloc excision remains the treatment of choice in all primary malignant chest wall neoplasms except large solitary plasmacytomas where incisional biopsy followed by irradiation appears to be the method of preference. In Ewing's and Askin's tumours, additional chemotherapy and radiotherapy have to be used. The extent of surgical excision should only be limited by the amount of tissue necessary to remove for adequate malignant tissue clearance, since even large defects can be reconstructed with little functional disturbance.

The treatment of patients with malignant superior vena caval obstruction with minimal morbidity has been made possible by the recent introduction of expandable metal stents as the sole palliative treatment or as an adjunct to other treatment modalities. To alleviate the distressing symptoms of superior vena caval obstruction, self-expanding metal stents were used successfully in 12 (Wallstent device in 6 and Gianturco device in 6 patients) of 13 patients. The diagnoses were small cell carcinoma ( n = 4), squamous cell carcinoma ( n = 4), non-Hodgkin's lymphoma ( n = 1), and mesothelioma ( n = 1), and a diagnosis of malignancy was not confirmed (although strongly suspected) in the remaining three cases. Eleven patients had immediate relief of obstruction and there was no change in one patient. Mean follow-up was 3.7 months (range 1 to 10 months). Excellent palliation was obtained in all but one patient in whom recurrent superior vena caval obstruction developed 3 months after stenting. Mean survival was 4.8 months (range 1 to 10 months). The ease of insertion with the use of local anesthesia with radiologic control, the self-expanding nature of the stent, and the lack of major complications on follow-up of up to 10 months are particular advantages. The self-expanding superior vena caval stents are a useful addition to our armamentarium in the management of malignant superior vena caval obstruction.(J Thorac Cardiovasc Surg 1996;112:335-40)

Traumatic diaphragmatic rupture remains a diagnostic challenge, and associated injuries determine the outcome in those diagnosed early, whereas that of latent cases is dependent on the consequence of the diaphragmatic rupture: namely, the diaphragmatic hernia. To analyze the clinical and radiologic features and the therapeutic implications, we reviewed 980 patients reported in the English-language literature. This injury affects predominantly males (male:female = 4:1) in the third decade of life, and is often caused by blunt trauma (75%). There were 1,000 injuries, of which 685 (68.5%) were left-sided, 242 (24.2%) right-sided, 15 (1.5%) bilateral, and 9 (0.9%) pericardial ruptures; 49 cases were unclassified. Chest (43.9%) and splenic (37.6%) trauma were the most common associated injuries. The diagnosis was made preoperatively in 43.5% of cases, whereas in 41.3% it was made at exploration or at autopsy and on the remaining 14.6% of the cases the diagnosis was delayed. The mortality was 17% in those in whom acute diagnosis was made, and the majority of the morbidity in the group that underwent operation was due to pulmonary complications. Uniform diagnosis depends on a high index of suspicion, careful scrutiny of the chest roentgenogram in patients with thoracoabdominal or polytrauma, and meticulous inspection of the diagphragm when operating for concurrent injuries. Repeated evaluation for days after injury is necessary to discern injury in patients not requiring laparotomy. Acute diaphragmatic injuries are best approached through the abdomen, as more than 89% of patients with this injury have an associated intraabdominal injury. Patients with diaphragmatic rupture presenting in the latent phase have adhesion between the herniated abdominal and intrathoracic organs, and thus the rupture is best approached via a thoracotomy.

Summary Paravertebral analgesia is highly effective in blocking unilateral afferent input from the trunk, but its bilateral use does not appear to have been described. Eight patients undergoing major abdominal vascular surgery had pre-operative bilateral paravertebral catheters inserted. A dose of 25 ml of bupivacaine 0.5% divided between the two catheters provided the basis of an effective intra-operative analgesic regimen which was continued postoperatively by infusion of bupivacaine for 4 days. Diclofenac and morphine completed a balanced analgesic regimen which was started pre-operatively and continued for 5 and 3 days respectively. Cardiovascular stability was notable throughout surgery, even with aortic clamping and all patients were extubated at the end of surgery. Good quality analgesia was obtained as assessed by low postoperative pain scores at rest and on movement. Three patients were not admitted to the intensive care unit, where mean stays for the group was 10 h (range 0-24 h) and mean hospital stay was 10 days (range 7–14). There were no additional demands made on the nursing staff and no postoperative pain-related complications occurred. We conclude that bilateral paravertebral analgesia, as part of a balanced analgesic technique, is an alternative method of providing effective afferent blockage for major abdominal surgery.

Six patients undergoing paravertebral blocks for chronically painful conditions of the chest wall were thermographically imaged so that the extent of cutaneous vasodilatation and hence sympathetic block could be correlated with the distribution of the somatic block. All blocks were performed by a single experienced operator, with a single percutaneous entry, using 15 ml of 0.5% bupivacaine at a mean level of T9-10 (range T7-8--T10-11), with radiological confirmation of correct needle placement. There was a mean distribution of the somatic block of five dermatomes (range 1-8), as evidenced by loss of pinprick sensation, with upper and lower limits of T6 and L3. The mean distribution of the sympathetic block was eight dermatomes (range 6-10), as evidenced by ipsilateral skin warming (p = 0.0005-0.001), with upper and lower limits of T5 and L3. No bilateral spread was observed. No significant postural changes in blood pressures were seen, although there was a small but significant decrease in supine heart rate (p = 0.05). This study demonstrates that a large unilateral somatic and sympathetic block is obtainable with a single thoracic percutaneous paravertebral injection. It challenges the suggestions that this method of analgesia is ineffective and hazardous, that a sympathetic component is a rare accompaniment and that the lumbar nerve roots are spared.

Plasma concentrations of bupivacaine have been measured in 12 patients given bupivacaine through a paravertebral catheter placed under direct vision at thoracotomy. After an initial bolus of 0.5% bupivacaine 20 ml, mean (SEM) CPmax was 1.45 (0.32) fig mt1 and median (range) tCPmax was 25 (10-60) min. A concentration of 4.43 ug mt' measured in one patient was not associated with toxic signs. During continuous infusion of bupivacaine for 120 h, CPmax was 4.9 (0.7) fig mt' and tCPmax 48 (5–96) h. No symptoms or signs of toxicity occurred. Separate measurement of R- and S-bupivacaine concentrations demonstrated significantly different concentration-time profiles. (Br. J. Anaesth. 1993; 70: 201–204)

Twenty patients undergoing elective thoracotomy were randomized into two groups, receiving either lumbar epidural morphine (n = 10) or continuous extrapleural intercostal nerve block (n = 10). Subjective pain relief was assessed on a linear visual analogue scale. Pulmonary function (peak expiratory flow rate, forced expiratory volume in 1 second, and forced vital capacity) was measured on the day before operation and daily for 4 days after operation. Pulse oximetry monitoring was used to determine the incidence of hypoxemia. No significant difference was observed between the groups concerning pain relief (except at 28 hours, in favor of the intercostal nerve block group), respiratory performance, or arterial oxygen saturation. Vomiting, pruritus, and urinary retention occurred only in the epidural group, whereas nausea occurred significantly less frequently in the extrapleural group. We conclude that after thoracotomy continuous extrapleural intercostal nerve block is as effective as lumbar epidural morphine in reducing postoperative pain and restoring pulmonary mechanics. Because of the significantly lower complication rates we favor continuous extrapleural intercostal nerve block for postthoracotomy analgesia.

One hundred eighty-seven consecutive patients underwent resection of primary bronchogenic carcinoma with intraoperative application of monomeric n-butyl-2-cyanoacrylate glue from July 1987 through December 1992. The glue reinforced either the stapled bronchial stump (135 patients), the sutured bronchial anastomosis in sleeve resections (37 patients) or the staple lines of wedge resections (15 patients). Mortality was 1.6% overall (3 of 187), and 5% among pneumonectomies (2 of 40). Bronchopleural fistulae occurred in 0.5% (1 of 187) of all pulmonary resections and 2.5% of pneumonectomies (1 of 40). There was no fistula in the lobectomy or sleeve resection groups. Bronchial anastomosis was accomplished in patients who underwent sleeve resection with four interrupted apposing sutures and airtight closure ensured by the tissue adhesive. There was no incidence of bronchial stenosis. There were no cyanoacrylate adhesive-related complications. A follow-up of the patients up to 68 months has indicated not only its effectiveness but also its safety. Monomeric n-butyl-2-cyanoacrylate glue is safe, offers protection to bronchial margins and may be valuable in preventing bronchial stenosis after sleeve resections.

Tuberculosis of the esophagus is rare. More cases of secondary tuberculosis of the esophagus have been reported compared to primary esophageal tuberculosis. An illustrative case of primary tuberculosis of the esophagus is presented and 54 published cases of esophageal involvement in tuberculosis are reviewed. There were eight primary and 46 secondary cases of esophageal tuberculosis. There were more male than female patients (M/F = 2.6:1) with an average age of 39.2 years (range 6 months to 75 years). Diagnosis can be difficult, being based on clinical, radiological, endoscopic, and histological features and on the response to chemotherapy. Most cases can be successfully treated with antituberculous chemotherapy even in the presence of fistulous tracts. Patients with esophageal tuberculosis in the presence of AIDS do not respond as well to antituberculous chemotherapy.

To assess the efficacy of continuous extrapleural intercostal nerve block (CEINB) with 0.5% bupivacaine on postoperative pain, pulmonary function and pulmonary complications, a prospective, randomised, double blind controlled trial was conducted on comparable patients undergoing lung resections; n = 21 (control), n = 25 (bupivacaine). The bupivacaine group required less papaveretum in the first 24 h (P less than 0.01) and had lower pain scores over 5 postoperative days (P less than 0.01). Pulmonary function recovered earlier in the bupivacaine group. Pulmonary complications occurred in 1 patient with normal lung function and 12 patients with obstructive airways disease (COAD): FEV1/FVC less than 70%. There were no infusion-related complications. CEINB has been shown to be safe and effective in reducing postoperative pain and pulmonary complications. CEINB minimises the loss of lung function after thoracotomy and restores impaired pulmonary mechanics more rapidly.

The major drawback of paravertebral space location using the traditional method of loss of resistance to air or saline is a failure rate of at least 10%. We investigated whether pressure measurement during needle advancement could improve reliability. Twenty-nine blocks in 14 awake adult patients undergoing treatment for chronically painful neuralgic conditions of the chest or abdominal wall were studied. In erector spinae, the mean inspiratory pressure (29.5 mmHg, SD 14.2), exceeded the expiratory pressure (19.4 mmHg, SD 9.7). However, upon traversing the superior costo-transverse ligament, there was a sudden lowering of pressures and the mean expiratory pressure (7.6 mmHg, SD 3.7) exceeded the inspiratory pressure (3.3 mmHg, SD 2.9). No negative pressures were recorded. Correct needle placement was confirmed by X ray screening and contrast injection. All blocks were successful and uncomplicated. Location of the paravertebral space by this objective method of 'pressure inversion' improves sensitivity and specificity and should lead to an improvement in the success rate of thoracic paravertebral analgesia.

Others have demonstrated the effectiveness of intercostal analgesia with bupivacaine hydrochloride (Marcain Plain; Astra). We present a greatly simplified method of effecting this. Our method is dependent on an intact pleura. To date, we have utilized this technique in 81 patients. Seventy-five (92.6%) required no additional analgesic in the first 24 hours following operation and 66 (81.5%), in the subsequent four days. Only 2 patients had postoperative pulmonary complications. No complication related either to the procedure or to the infusion of bupivacaine occurred. The technique as described here is a safe and reliable method of providing analgesia without any side effects after thoracotomy.

Three cases of primary malignant melanoma of the oesophagus are presented and the literature is briefly reviewed. This rare disease affects predominantly males in the sixth decade of life or later and the clinical and radiologic features are indistinguishable from those of more common oesophageal carcinomas. The tumours are often polypoid, located mainly in the lower two-thirds of the esophagus, and their histologic features resemble those of lentigo maligna melanoma, but with more aggressive biologic behavior. Because of the potential for extensive intramucosal involvement, multicentricity and high local recurrence rate, surgical treatment requires radical excision with a much greater margin than for the common squamous cell carcinomas. Though the outlook is poor, surgery is the favoured treatment with palliative or curative intent, with a 5-year survival rate of 4.2%. Radiotherapy, chemotherapy and immunostimulation currently serve mainly as palliative or adjunctive measures.

Surgical closure of a bronchopleural fistula following right pneumonectomy for squamous cell carcinoma was achieved with monomeric n-butyl-2-cyanoacrylate. Previous attempts at closure by conservative and endoscopic means were unsuccessful. Tissue adhesives are useful in the surgical or endoscopic management of bronchopleural fistula.

Thoracic paravertebral nerve blockade, although once widely practised, has now only a few centres which contribute to the literature. Data production has, however, continued and this review correlates this new information with existing knowledge. Its history, taxonomy, anatomy, indications, techniques, mechanisms of analgesia, efficacy, contraindications, toxicity, side effects and complications are reviewed. Thoracic paravertebral analgesia is advocated for surgical procedures of the thorax and abdomen, especially wherever the afferent input is predominantly unilateral eg. thoracotomy, cholecystectomy and nephrectomy. It is also of benefit in the prevention and management of chronic pain. It is a simple undertaking with impressive efficacy. Plasma local anaesthetic levels are acceptable and its side effect and complication rates are low. No mortality has been reported. For unilateral surgery of the chest or trunk, thoracic paravertebral analgesia should be considered as the afferent block of choice. For bilateral surgery, its efficacy may be limited by the doses of local anaesthetic which could safely be used and further study in this area in particular is required. This form of afferent blockade deserves greater consideration and investigation.

We have undertaken a prospective, randomized comparison of the superficially similar techniques of interpleural and paravertebral (extrapleural) analgesia in 53 patients undergoing posterolateral thoracotomy. Local anaesthetic placed anterior to the superior costotransverse ligament and posterior to the parietal pleura produces a paravertebral block and instilled between the parietal and visceral pleurae produces an interpleural block. Patients received preoperative and postoperative continuous bupivacaine paravertebral blocks in group 1 and interpleural blocks in group 2. Premedication comprised diclofenac and morphine, and after operation all patients had regular diclofenac and patient-controlled morphine (PCM). Analgesia was assessed by visual analogue pain scores (VAS), PCM requirements, ratio of preoperative to postoperative spirometric values (PFT), rates of postoperative respiratory morbidity (PORM) and hospital stay, all recorded by blinded observers. Eight patients were withdrawn and data from 45 patients were analysed. Patient characteristics, surgery, VAS scores and PCM use were similar in both groups. PFT were significantly better (P = 0.03-0.0001) in group 1, and PORM was lower and hospital stay approximately 1 day less in this group. Five patients in group 2 became temporarily confused, probably because of bupivacaine toxicity (P = 0.02). We conclude that bupivacaine deposited paravertebrally produced greater preservation of lung function and fewer side effects than bupivacaine administered interpleurally.

Esophageal fistula formation after pulmonary resection is rare. We present a case of esophagocutaneous fistula 9 years after left pneumonectomy for squamous cell carcinoma of the bronchus. This was confirmed on barium swallow and endoscopy. The fistula was sealed successfully with monomeric n-butyl-2-cyanoacrylate tissue adhesive. The simplicity and effectiveness of this method of management is discussed and the multitude of surgical options considered.

The role of surgery in the management of Stage III lung cancer is controversial. A retrospective analysis of our experience with 220 Stage III patients treated surgically (Stage IIIa n = 174, Stage IIIb n = 46) form the basis of this report. Of the 140 patients who underwent resection only 88 were considered to be potentially curative, all but two being operated in Stage IIIa. The overall 5-year survival rate for Stage III lung cancer was 12.1% whilst curative resection had a 5-year survival rate of 31.8%. The 5-year survival rate increased to 42% with curative resection for T3N0M0 patients. There were no 5-year survivors with incomplete resection. There were no 5-year survivors in Stage IIIb disease. Five-year survival rates for N0 (n = 62), N1 (n = 78), and N2 (n = 80) irrespective of T status were 37.2%, 3.4% and 5.6%, respectively. The operative mortality rate was only 3.4% for curative resection while palliative or non-resection patients had a mortality rate of 10.6%. Eighty-eight patients, 55 of Stage IIIa and 23 of Stage IIIb had postoperative radiotherapy with a median survival of 12 and 9 months, respectively. Fifteen small cell carcinoma patients had postoperative adjuvant chemotherapy with a median survival of 6.5 months. The only 5-year survivor in this group also had a curative resection. We conclude that metastasis to lymph nodes usually implies systemic disease and a poor prognosis. Surgical therapy continues to be the treatment of choice in a small subset of patients with Stage III resectable lung cancer discovered at thoracotomy.(ABSTRACT TRUNCATED AT 250 WORDS)

Obstruction of the superior vena cava was relieved in two patients by intraluminal, self-expanding Gianturco stents percutaneously inserted in local anaesthesia. The cause of the obstruction was inoperable oat cell carcinoma of the bronchus in one case and irradiation for squamous cell carcinoma of the right upper lobe in the other case. Stenting gave immediate relief of symptoms in both cases, with no complications. Intraluminal stenting is a useful adjunct in the management of patients with obstruction of the superior vena cava.

Six children with a mean age of 10.6 years (range, 7 to 16 years) underwent thoracotomy for pulmonary and esophageal procedures. Postoperatively, continuous paravertebral block using an infusion of bupivacaine via an extrapleural catheter was used. Excellent analgesia was attained in all patients, with no requirement for opiates or other analgesic drugs. There were no pulmonary complications and no complications related to the continuous extrapleural infusion. We conclude that continuous paravertebral block is an effective and safe method for postthoracotomy pain relief in children.

To evaluate the effects of continuous extrapleural intercostal nerve block on post-thoracotomy pain and pulmonary complications, a randomized, double-blind, placebo-controlled study was conducted on 80 patients undergoing elective thoracotomy for pulmonary (n = 47) or oesophageal (n = 33) procedures. In patients who received continuous bupivacaine infusion, the requirement for intramuscular opiate and rectal diclofenac was less, the score on a visual linear analogue pain scale lower and recovery of pulmonary function more rapid than in saline-infused controls. Postoperative pulmonary complications occurred in 35% of the saline group, but only 10% of the patients with bupivacaine infusion (p0.01). In patients with chronic obstructive airways disease (COAD), the incidence of postoperative pulmonary complications was 54.5% in the saline group and only 4.5% in the bupivacaine group (p0.01). Among the patients without COAD there was no significant intergroup difference in such complications. We conclude that continuous extrapleural intercostal nerve block is effective for post-thoracotomy analgesia and reduces pulmonary complications of thoracotomy in patients with COAD.

A previously unreported coupled achiral-chiral high-performance liquid chromatographic method has been developed to assay the levels of bupivacaine and its enantiomers in plasma samples, after the local anaesthetic had been given as a continuous extrapleural intercostal nerve block for 4 days, to relieve postoperative pain following thoracotomy. The method has been used to determine maximum, individual enantiomer and steady state levels in conjunction with an assessment of whether accumulation of bupivacaine occurs. An off-line extraction sample preparation is involved before determination of the "total" levels and final sample clean-up on a cyanopropyl silica column prior to "heart cutting" of the bupivacaine peak to a Chiral-AGP column for assay of the enantiomeric ratio. For an initial 5 patient number the mean maximum level was 5.43 micrograms/ml against a reported toxic level of around 5 micrograms/ml, which was reached in 72 h and the S-enantiomer gave slightly increased concentrations over the R-enantiomer for which there is some evidence of higher toxicity.

To analyze the importance of airway involvement in relapsing polychondritis, an illustrative case report is presented and 62 patients reported in the literature with serious airway complications are reviewed. There were 47 female and 17 male patients, with an average age of 40.3 years (range, 2 to 73 years). Patients were seen with hoarseness, breathlessness, cough, stridor, wheezes, and tenderness over laryngotracheal cartilages. Respiratory tract involvement was confirmed by conventional radiography, tomography, computed tomography, dynamic pulmonary function tests, and bronchoscopy. Corticosteroids and antiinflammatory and immunosuppressive agents were used in these patients. Tracheostomy was performed in 18 patients. Death occurred in 13 patients despite tracheostomy or corticosteroid therapy, or both. A detailed analysis of the clinical, radiological, and pulmonary function studies is presented, with emphasis on upper airway mechanics. The medical and surgical management options are reviewed, including the use of endotracheal prosthesis and extraluminal splinting in dynamic airway collapse.

Continuous extrapleural intercostal nerve block has been shown in a randomized, controlled study to be effective in reducing postoperative pain after thoracotomy and in restoring pulmonary mechanics. To assess the extent of spread of bupivacaine infused through an extrapleurally placed cannula inserted at thoracotomy, iohexol (Omnipaque) was infused at 5 days postoperatively in 5 patients and computed tomography performed. The contrast medium was confined to the paravertebral space covering on average six intercostal spaces. This study demonstrated that anatomically, the site of action of the bupivacaine infused through an extrapleural cannula was primarily in the paravertebral space.

A randomised, double blind trial was carried out in 16 patients undergoing pleurectomy to assess the effect of continuous extrapleural intercostal block on postoperative pain and pulmonary function. Subjective pain relief was assessed on a linear visual analogue scale. Pulmonary function was measured on the day before operation and daily for five days after surgery. Eight patients received bupivacaine and eight placebo (saline). The mean pain scores at 4, 8, 16, and 24 hours were 13.3, 8.5, 6.1, and 10 mm respectively in the bupivacaine group compared with 56.3, 41, 46.7, and 35 in the control group; in addition, the bupivacaine group required less papaveretum. Twenty four hours after surgery mean values of peak expiratory flow, forced expiratory volume in one second, and forced vital capacity were reduced to 82%, 76%, and 76% of preoperative control values in the bupivacaine group, and to 39%, 32%, and 36% in the control group. The speed of recovery of pulmonary function was superior in the bupivacaine group. There were no complications related to the infusion. Continuous extrapleural intercostal nerve blockade with bupivacaine provides safe and effective postoperative analgesia and improves respiratory mechanics after pleurectomy.

Monomeric n -butyl-2-cyanoacrylate was used in 25 patientsundergoing pulmonary resections to strengthen the bronchial stump after pneumonectomy (n = 11) and to aid bronchial (n = 13) and tracheal anastomosis (n = 1) after sleeve resections. Neither group had any incidence of bronchopleural fistula. Bronchial anastomosis was accomplished in patients who underwent sleeve resection, reducing the number of sutures required to four apposing sutures, with the tissue adhesive ensuring an airtight closure. There was no incidence of bronchial stenosis. The efficacy of n -butyl-2-cyanoacrylate in preventing fistula formation after bronchial resections makes it an ideal agent in pulmonary surgery. Its use obviated the use of pedicled pleural flap, thus ensuring pleural integrity for extrapleural continuous intercostal nerve blockade for postoperative analgesia.

An unusual case of carcinosarcoma of the lung with metastasis to the gastrointestinal tract causing intussusception is presented. The clinical, radio-logic and histopathologic features are reviewed with reference to relevant literature.

A case of spontaneous rupture of the oesophagus (Boerhaave's syndrome) following acute myocardial infarction is presented. Conservative management with chest drainage and parenteral, later jejunostomy, feeding was successful. Progress was monitored with methylene blue swallow. A review of the literature revealed only two previous cases of this type.

To assess the efficacy of continuous extrapleural intercostal nerve block on postoperative pain and pulmonary function, a prospective, randomized, double-blind, placebo-controlled trial was conducted on 56 patients undergoing elective thoracotomy. Infusion was started before closing the chest and was continued for 5 days. Subjective pain relief was assessed on a linear visual analogue scale. Pulmonary function was measured on the day before operation and daily for 5 days. There were 29 patients in a group which received bupivacaine and 27 in a control group which received saline. The bupivacaine group had lower pain scores (P < 0·01) and required less papaveretum (P < 0·01) than the control group. Forced vital capacity, forced expiratory volume in 1 s and peak expiratory flow rate were maximally reduced at 24 h to median values of 56, 60 and 57 per cent, respectively, of preoperative control values in the bupivacaine group, and to 25, 30 and 32 per cent in the control group. These differences are highly significant (P < 0·01). Restoration of pulmonary function was superior in the bupivacaine group (P < 0·01). There were no infusion-related complications. After thoracotomy, continuous intercostal blockade with bupivacaine is a safe and effective method of pain relief which reduces the early loss of postoperative pulmonary function significantly and more rapidly restores respiratory mechanics.

Twelve cases of oesophageal intramural pseudodiverticulosis are described and the findings in 85 previously reported cases are reviewed. The condition occurs in all age groups, predominantly in the sixth and seventh decades, with a slight predilection for males. The characteristic radiographic appearance is of multiple flask shaped outpouchings of 1-4 mm with narrow necks communicating with the oesophageal lumen. The source of the pseudodiverticula has been shown to be pathologically dilated excretory ducts of the submucous glands due to chronic submucosal inflammation. The distribution was segmental in 57 cases (59%) and diffuse in 40 (41%). Dysphagia is the main symptom and was found in 85 cases (88%); 88 cases out of 97 had radiological narrowing of the oesophagus; of these, 39 (44%) were in the upper oesophagus, 20 (23%) in the middle oesophagus, and 29 (33%) in the lower oesophagus. Treatment is directed towards management of the associated disorder, as the diverticula themselves rarely cause problems.

Nine cases of primary sternal tumour were retrospectively reviewed in regard to clinical, radiologic and surgical features. The tumours were chondrosarcoma (4), chondroma (3), solitary plasmacytoma (1) and osteochondroma (1). All the benign lesions were excised without complication or recurrence. Three chondrosarcomas were radically excised en bloc, and one was only locally excised because of gross involvement of underlying structures. The solitary plasmacytoma was treated with incisional biopsy and radiotherapy. Two of the three patients with radical excision of sternal chondrosarcoma were alive after 5 years. Careful preoperative assessment, including use of computed tomography, is important. Wide excision should be the procedure of choice for all sternal tumours, since differentiation between benign and malignant lesions may be difficult in cartilaginous tumours, which are the commonest types.

A retrospective study of 53 primary chest wall tumors, 26 benign and 27 malignant, was carried out to review their clinical radiological and pathological features. Forty-nine of the 53 lesions presented in the ribs and the remaining 4, in the sternum. The overall 5-year survival for patients with primary malignant neoplasms of the chest wall was 33.3%, and the 10-year survival was 18.5%. All of the deaths were disease related. All of the patients with benign tumors were treated by excision without recurrence or death. Distinction between benign and malignant chest wall tumors was not possible using radiographic criteria unless cortical destruction and involvement of soft tissues were visualized. On the basis of our analysis, we believe that all tumors of the chest wall should be considered malignant until proven otherwise and that wide excision should be carried out. This is necessary not only to obtain an adequate diagnosis but also to provide the best chance for cure in both benign and malignant lesions.

Aneurysmal bone cysts are uncommon lesions, especially in the ribs. Four patients with aneurysmal bone cysts of the rib are presented and previously reported cases reviewed. A brief discussion of the clinical manifestations, pathology, aetiology, and current treatment of aneurysmal bone cyst is also included.

A rare case of metastatic malignant melanoma of the esophagus 11 years after wide excision of a cutaneous malignant melanoma is presented and the relevant literature reviewed. Symptomatic esophageal metastasis is unusual in malignant melanoma. Although the eventual outlook is poor in such cases, worthwhile palliation can be achieved by prompt diagnosis and treatment.