Significance of the number, size and type of drains used for intercostal drainage after oesophagectomy for cancer

Abstract Background Pleural space drainage with intercostal drains (ICD) is performed after oesophagectomy to allow the lung to reinflate, remove excess fluid post-operatively, and signal chyle or enteric content. Enhanced recovery protocols encourage the use of the minimum number of drains for the shortest duration to facilitate rapid recovery after surgery. There is wide variability in the type, number and size of drains inserted at operation. This study sought to identify the most effective drain pattern insertion, using the need for respiratory reintervention as the primary end point and secondary outcome of the presence of pleural effusions. Methods All patients undergoing oesophagectomy for cancer in one unit were included between November 2014 and December 2020. The operation performed, drain sizes, sides and type were recorded. Respiratory reintervention was defined as replacement of an ICD, bronchoscopy, pleural aspiration or reintubation. The primary and secondary end points, and potential confounders such as age, histology, pre-operative stage of disease, neoadjuvant therapy, pre-existing lung disease, and anastomotic or chyle leak were recorded. Results The study period encompassed 258 patients who underwent oesophagectomy for cancer. Median age 69 (range 32-82), 211 male, 226 ACA:32 SCC, 224 neoadjuvant therapy, 212 right-sided thoracic operations, 46 left thoracoabdominal approach. Post-operative respiratory reinterventions occurred in 47 patients (18.2%). At least one post-operative pleural effusion was present in 52 patients (20.2%): 9 bilateral; 26 contralateral; 17 ipsilateral to the side of thoracic surgery. 67% of effusions were contralateral to the operated side. The use of two or three ICDs (HR 371683269, p < 1), one or two operative side ICDs (HR 0, p < 1), Blake’s drains in place of rigid ICDs (HR 0.938 [0.422-2.085], p < 0.875), and size 24F compared to 28F drains (HR 0, p < 0.999) are not significantly associated with post-operative respiratory reinterventions. Similarly, the presence of post-operative pleural effusions is not significantly associated with the use of two or three ICDs (HR 240242843, p < 1), one or two operative side ICDs (HR 0, p < 1), Blake’s drains in place of rigid ICDs (HR 1.505 [0.665-3.405], p < 0.327), and size 24F compared to 28F drains (HR 1.055 [0.109-10.2], p < 0.963). Conclusions This study supports the use of contralateral pleural space drainage as two thirds of effusions were contralateral to the operated side. It shows no correlation between the size of drains, number of drains or use of Blakes drains and the likelihood of requiring a post-operative respiratory intervention or development of post-operative pleural effusion. Therefore the ERAS principles of the fewest number of drains for the shortest duration should be adopted.