The incidence and management of postoperative chylothorax after pulmonary resection and thoracic mediastinal lymph node dissection

Abstract

Background: Our objective was to determine the incidence and optimal management of chylothorax after pulmonary resection with complete thoracic mediastinal lymph node dissection (MLND).

Methods: This is a retrospective review of patients who underwent pulmonary resection with MLND.

Results: Between January 2000 and December 2012, 2,838 patients underwent pulmonary resection with MLND by one surgeon (RJC). Forty-one (1.4%) of these patients experienced a chylothorax. Univariate analysis showed that lobectomy (p<0.001), a robotic approach (p=0.03), right-sided operations (p<0.001), and pathologic N2 disease (p=0.007) were significantly associated with the development of chylothorax. Multivariate analysis showed that lobectomy (p=0.011), a robotic approach (p=0.032), and pathologic N2 disease (p=0.027) remained predictors. All patients were initially treated with cessation of oral intake and 200 μg subcutaneous somatostatin every 8 hours. If after 48 hours the chest tube output was less than 450 mL/day and the effluent was clear, patients was given a medium-chain triglyceride (MCT) diet and were observed for 48 hours in the hospital. If the chest tube output remained below 450 mL/day, the chest tube was removed, they were discharged home with directions to continue the MCT diet and to return in 2 weeks. Patients were instructed to consume a high-fat meal 24 hours before their clinic appointment. If the patient's chest roentgenogram was clear at that time, they were considered "treated." This approach was successful in 37 (90%) patients. The 4 patients in whom the initial treatment was unsuccessful underwent reoperation with pleurodesis and duct ligation.

Conclusions: Chylothorax after pulmonary resection and MLND occurred in 1.4% of patients. Its incidence was higher in those with pathologic N2 disease and those who underwent robotic resection. Nonoperative therapy is almost always effective.