Nutritional predictors of postoperative outcome in pancreatic cancer (Br J Surg 2011; 98: 268–274)

Sir We read with great interest the recent study by Kanda and colleagues, which seeks to establish a relationship between preoperative ‘nutritional status’ and survival. However, we would appreciate clarification of some of the issues raised. The components of Onodera's prognostic nutrition index—serum albumin and total lymphocyte count—correlate with increased morbidity and mortality after a variety of types of surgery. However, we wonder whether, either individually or together, these markers can be said to reflect nutritional status accurately. Although inadequate nutrition in advanced pancreatic cancer may contribute to low serum albumin levels, given the marked acute-phase response present in this patient population and albumin's role as a negative acute-phase serum protein1, the measurement cannot distinguish between albumin as a marker of nutritional status or aggressive cancer phenotype. The same is true of total lymphocyte count2. The prognostic value of body mass index (BMI) was not statistically significant. The relationship between BMI and nutrition is complex in pancreatic cancer. For instance, in North America the average BMI at presentation for palliative management of advanced cancer is now 26 kg/m2 (V. Baracos and L. Martin, personal communication), and the cachectic process continues underneath this ‘mantle’ of adiposity. Thus patients who have a high rather than a low BMI can have sarcopenia, and a poor prognosis3. This demonstrates that BMI is a limited marker of nutritional status and may explain the null result. Some exclusion criteria in the study were explained inadequately. Patients were recruited over an 18-year interval, over which there have been major advances in perioperative management and nutritional/metabolic care. Indeed, low BMI is no longer an independent risk factor for adverse outcome after major abdominal surgery when an enhanced recovery after surgery approach to perioperative care is used4. Patients with no preoperative oral intake were also excluded. One would expect patients in this subgroup to be severely malnourished and thus provide insight into the importance of preoperative nutritional status. The latter patients are common, and their exclusion limits the relevance of the results to clinical practice.