Propensity score-based analysis of outcomes of laparoscopic versus open liver resection for colorectal metastases.

ABSTRACT

BACKGROUND: There is a need for high-level evidence regarding the added value of laparoscopic (LLR) compared with open (OLR) liver resection. The aim of this study was to compare the surgical and oncological outcomes of patients with colorectal liver metastases (CRLM) undergoing LLR and OLR using propensity score matching to minimize bias. METHODS: This was a single-centre retrospective study using a prospective database of patients undergoing liver resection for CRLM between August 2004 and April 2015. Co-variates selected for matching included number and size of lesions, tumour location, extent and number of resections, phase of surgical experience, location and lymph node status of primary tumour, perioperative chemotherapy, unilobar or bilobar disease, synchronous or metachronous disease. Prematching and postmatching analyses were compared. Surgical and oncological outcomes were analysed. RESULTS: Some 176 patients undergoing LLR and 191 having OLR were enrolled. After matching, 133 patients from each group were compared. At prematching analysis, patients in the LLR group showed a longer overall survival (OS) and higher R0 rate than those in the OLR group (P = 0·047 and P = 0·030 respectively). Postmatching analyses failed to confirm these results, showing similar OS and R0 rate between the LLR and OLR group (median OS 55·2 versus 65·3 months respectively, hazard ratio 0·70 (95 per cent c.i. 0·42 to 1·05; P = 0·082); R0 rate 92·5 versus 86·5 per cent, P = 0·186). The 5-year OS rate was 62·5 (95 per cent c.i. 45·5 to 71·5) per cent) for OLR and 64·3 (48·2 to 69·5) per cent for LLR. Longer duration of surgery, lower blood loss and morbidity, and shorter postoperative stay were found for LLR on postmatching analysis. CONCLUSION: Propensity score matching showed that LLR for CRLM may provide R0 resection rates and long-term OS comparable to those for OLR, with lower blood loss and morbidity, and shorter postoperative hospital stay.