A prospective, randomised self controlled trial with the approval of the East Birmingham Research Ethics Committee was performed at Birmingham Heartlands Hospital between December 2005 and October 2007. We used the CONSORT checklist for design and conduct of this study. As there are no comparative trials comparing these interventions we designed the study in two phases a pilot phase of ten patients to review the results and to calculate a sample size for the trial. An informed patient consent was obtained before each operation. Patients undergoing bilateral LVRS through a median sternotomy incision were included in the study. Patients with asymmetrical disease and known allergies to bovine pericardium and albumin were excluded. All patients had routine work-up for LVRS according to our unit protocol, comprising full lung function tests, high resolution CT scanning (HRCT), quantitative ventilation-perfusion scanning, echocardiography in patients with previous cardiac history, smoking cessation and pulmonary rehabilitation.

Randomisation was undertaken with sequential closed envelopes containing the treatment strategies assigned to each side. In each case, the operating surgeon opened the envelope on the day of surgery to assign the treatment strategy to each side (Figure 1).

All patients had a peri-operative epidural catheter sited for pain relief. The surgery was performed under general anaesthesia with a double lumen endotracheal tube and sequential single lung ventilation. The surgical access was gained through median sternotomy. The target areas were identified based on preoperative CT scan, ventilation perfusion scan as well as by direct per-operative observation and palpation of lung parenchyma. The pleural cavity was entered after instituting single lung ventilation to the contra lateral side. After few minutes of suspending ventilation, the relatively better part of the lung parenchyma tends to collapse thus demarcating the worst areas for excision. The line of excision began at the medial aspect, near the horizontal fissure for the right upper lobe and near the base of the lingula for the left upper lobe. With successive applications of the GIA 80 stapler (Auto Suture, Tyco Healthcare Norwalk CN, USA), the line of excision was carried up toward the apex, angled postero-laterally, and then angled downward on the postero-lateral portion of the upper lobe ending up near the top of the oblique fissure. Thus an inverted 'U' shaped continuous staple line was formed. The resection was limited to the level of the azygos vein on the right and to the aortic arch on the left.

BioGlue was applied on the stapled margins using a spreader tip. The unique double-helix delivery system attached to a syringe enables swift mixing of two components of the product. The lung was gently re-inflated after two minutes. In the other group Peri-strips were applied onto the jaws of the GIA stapler before applying them on the lung parenchyma.

After ensuring pneumostasis, 28F apical and basal drains were placed in the pleural space and brought out through separate stab incisions. The pleural openings on either side were closed with a continuous run of 4/0 prolene suture (Ethicon USA) to prevent crossover air leak. Once the pleurae were isolated the air-leak check was repeated to rule out crossover air leaks.

The anaesthetist carefully avoided over-inflation of the operated lung letting the lung inflate gently accepting a period of relative hypercarbia. At the end of the procedure, the sternum was closed with sternal wires and the pre-sternal fascia and skin were closed in layers. The patients were all extubated at the end of the operation and nursed in the thoracic high dependency unit.

The chest drains were connected to underwater seal drainage systems individually and left without suction. The drains were connected to flutter valve bags (Portex Ltd, Hythe UK) once the drainage was less than 100 mls/day to enable mobilisation. The chest drains were removed following cessation of the air leak, confirmation of full expansion of the lung and drainage less than 100 ml in 24 hours. This was achieved in majority of patients. However, if there was persistent air leak or the lung failed to expand after seven days, the patient was discharged home with drain connected to a flutter valve bag, which was removed subsequently. The patients were routinely followed up to 6 weeks after surgery.

In this study, the postoperative air leak was defined as the presence of air bubbles in the chest drains during the course of normal or forced expiration (coughing). The investigators performed daily objective assessments that were concurrently verified by independent blinded senior nursing staff.

The results are expressed as mean ? standard deviation. The sides randomised to each arm were grouped according to treatment allocation, compared with t tests for normally distributed measures and Mann-Whitney tests for non-normally distributed measures.

A total of 10 patients were recruited and consented over a period of 24 months into this study. There were 6 men with a mean age of 59.8 ? 4.9 years. All patients had discontinued smoking prior to surgery. The median Karnofsky performance scale of this group was 70%. Co-morbidities included hypertension (n: 1), diabetes mellitus (n: 1) and aortic regurgitation (n: 1). There was no significant difference in the distribution of emphysema between the two treatment sides as assessed by independent review of HRCT and quantitative ventilation perfusion imaging. A summary of pre-operative investigations is presented in Table 1.

There was shorter mean duration of air leak, less drainage and chest drains were removed earlier in the BioGlue group but these do not reach statistical significance due to small numbers (Figure 2, Table 2). There was a tendency for earlier cessation or comparable duration of air leak in the BioGlue arm on direct comparison between the two treatment sides in all but one patient. Two patients did not have any post operative air leak. BioGlue treated sides had comparable duration of air leak with Peri-strips treated sides. In three patients the BioGlue treated side had significantly shorter air-leak duration and one patient had a longer air-leak in the BioGlue treated side (Figure 3).

When comparing the number of subjects with air-leak on each of the post operative days there was only one patient in the BioGlue treated side who had prolonged air leak (duration more than 7 days) compared with four in the Peri-strips treated side (Figure 4).

There was single mortality on the 38^th post operative day, due to postoperative respiratory failure leading to multi-organ failure.

No adjunct related complications were encountered in the BioGlue arm. In Peri-strips side, one patient had air leak at the end of the procedure from the junction of two staple lines and required extra pneumostasis with insertion of prolene sutures. Although this can be viewed as technical failure, it may still be considered as adjunct failure. BioGlue application on the staple confluence may have sealed the leak where as with the Peri-strips it is incorporated into the overlapped stapling lines therefore necessacitating extra sutures. One patient had prolonged air leak on the Peri-strips treated side and was discharged home on a flutter valve bag. Another patient coughed up part of the Peri-strips three months after the surgical procedure.

This study was supported by an educational grant from Cryolife Europa. This work was presented in the Annual meeting of the Society for Cardiothoracic Surgery in Great Britain and Ireland in March 2008 and the 16th European Conference on General Thoracic Surgery in Bologna in June 2008.