Suction on chest drains following lung resection: evidence and practice are not aligned

A best evidence topic in Interactive CardioVascular and Thoracic Surgery (2006) looked at application of suction to chest drains following pulmonary lobectomy. After screening 391 papers, the authors analysed six studies (five randomized controlled trials [RCTs]) and found no evidence in favour of postoperative suction in terms of air leak duration, time to chest drain removal or length of stay. Indeed, suction was found to be detrimental in four studies. We sought to determine whether clinical practice is consistent with published evidence by surveying thoracic units nationally and performing a meta-analysis of current best evidence.We systematically searched MEDLINE, EMBASE and CENTRAL for RCTs, comparing outcomes with and without application of suction to chest drains after lung surgery. A meta-analysis was performed using RevMan(©) software. A questionnaire concerning chest drain management and suction use was emailed to a clinical representative in every thoracic unit.Eight RCTs, published 2001-13, with 31-500 participants, were suitable for meta-analysis. Suction prolonged length of stay (weighted mean difference [WMD] 1.74 days; 95% confidence interval [CI] 1.17-2.30), chest tube duration (WMD 1.77 days; 95% CI 1.47-2.07) and air leak duration (WMD 1.47 days; 95% CI 1.45-2.03). There was no difference in occurrence of prolonged air leak. Suction was associated with fewer instances of postoperative pneumothorax. Twenty-five of 39 thoracic units responded to the national survey. Suction is routinely used by all surgeons in 11 units, not by any surgeon in 5 and by some surgeons in 9. Of the 91 surgeons represented, 62 (68%) routinely used suction. Electronic drains are used in 15 units, 10 of which use them routinely.Application of suction to chest drains following non-pneumonectomy lung resection is common practice. Suction has an effect in hastening the removal of air and fluid in clinical experience but a policy of suction after lung resection has not been shown to offer improved clinical outcomes. Clinical practice is not aligned with Level 1a evidence.