Background Despite the positive results of the NETT trial in favour of lung volume reduction surgery (LVRS) uptake of the technique has been limited largely due to an exaggerated fear of the associated mortality risk. We have analysed our 18 year experience of LVRS to provide a more sophisticated personalised risk profile based on individual patient data.

Methods Since 1994 we have performed 250 lung volume reduction procedures on 220 patients: 153M:97F, age 61 (39–74) years. The initially approach was through median sternotomy (20 patients), with the subsequent 230 procedures performed by video-assisted thoracoscopic surgery (VATS), 3 of which required conversion to open thoracotomy. All patients underwent standard physiological and anatomical selection techniques with 51 (20%) falling outside recognised safety limits (FEV1 or DLCO <20% predicted). All patients were offered surgery after discussion in our LVRS MDT panel and counselled on risk on their basis of their physiological status. We analysed data collected prospectively using logistic regression to identify the factors predicting early postoperative mortality.

Results Open surgery significantly increased the risk of 30 day mortality 22% vs VATS 3.6% (p=0.005). Bilateral vs unilateral VATS had no influence. At 30 days mortality was associated with low BMI, DLCO and KCO At 90 days, mortality was also associated with FEV1 and RV:TLC DLCO was the only significant independent predictor of 30 day (OR 0.88, CI 0.80–0.97) and 90 day (OR 0.92, CI 0.88–0.98) mortality after VATS (table 1).

The causes of death after 30 days in the VATS group were mainly due to pneumonia (5 cases) with cardiac complications (2); tension pneumothorax (1) and fatal pulmonary haemorrhage (1) in the remainder.

Conclusion LVRS is primarily a procedure to improve health status so accurately informed consent is imperative. Careful consideration of preoperative physiological characteristics and operative technique allows estimation of an individualised mortality risk for LVRS which may be lower than the commonly perceived overall figure.