Preoperative determinants of the length of hospital stay in patients undergoing surgery for pancreatic cancer

Introduction: Surgery for pancreatic cancer is associated with considerable morbidity. Both surgeon-related and patient-related factors are important in determining postoperative outcome. Therefore most pancreatic surgery is now carried out in specialist centres. However, patient optimisation is complex since factors that determine postoperative morbidity are less well understood. Objective: To compare preoperative measures of patient comorbidity with length of hospital stay in patients undergoing pancreatic surgery. Patients and methods: 103 patients who underwent surgery for pancreatic cancer (pancreatico-duodenectomy n=87,trial dissection with bypass n=16) had preoperative evaluation of body habitus, systemic inflammatory response (Glasgow Prognostic Score, mGPS) and routine blood analysis. Cardiopulmonary function (CPET) was measured in 54 patients. Length of hospital stay was recorded. Results: The median length of hospital stay was 17 days. 39 stayed ≤ 14 days and 64 >14 days. Length of hospital stay was not associated with body habitus or systemic inflammatory response. However, in patients with CPET, more patients with low anaerobic threshold (p=0.030),high base excess (0.056) and high VE/VCO2 (p=0.022) had a hospital stay >14 days. Prolonged hospital stay was associated with pancreatico-duodenectomy (p=0.027) and anastomotic failure (p Conclusions: CPET was the only pre-operative objective measure that correlated significantly with prolonged hospital stay. Length of stay is a useful surrogate marker of post-operative adverse events. Although CPET does not predict specific complications, it may determine the host response to major surgery and its immediate sequelae.