11261-5614’89 OOOR+3197~$10.~lO CLINICAL NUTRITION (1989) 8: 197-201 zyxwvuts i zyxwvutsrqponmlkjihgfedcbaZ Longman Group UK Lrd 1989 zyxwvuts Prognostic Ability of Nutritional Assessment Methods in Surgical Cancer Patients zyxwvutsrqponmlkjihgfedcbaZYXWVUT M. Braga*, P. Baccari*, G. Radaellif, V. DiCarlo*, L. Gianotti* and M. Cristallo* *Scientific Institute San Raffaele Hospital, Cattedra Patologia Chirurgica-University of Milan, Via Olgettina 60,20132 Milan, tInstitute of Biometry and Medical Statistics, University of Milan, Italy. ‘Reprint requests to M.B.) ABSTRACT To identify patients at high-risk for post-operative infections, several methods have been proposed, including prognostic nutritional index (PNI), instant nutritional assessment (INA) and nutritional assessment (NA). Weight loss (WL) has also been related to post-operative morbidity. We have evaluated the prognostic ability of PNI, INA, NA and WL in a prospective study carried out in 94 patients affected by gastro-intestinal malignancy, who underwent major surgery. Post-operative infections occurred in 26 (27.70,a) patients. PNI, INA and NA identified classes of patients with a progressive risk of septic complications. To determine the prognostic ability of PNI, INA, NA and WL, sensitivity, specificity, Youden index and predictive values were evaluated. All methods,had a Youden index greater than one, with a positive predictive value ranging from 0.33 to 0.36. Since all the methods studied showed a similar predictive ability, it seems reasonable to identify the high-risk surgical patient by using weight loss in association with those nutritional parameters derived from routine hospital laboratory tests. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA INTRODUCTION Malnutrition is a possible cause of post-operative septic complications in the surgical patient [l-5]. To identify high risk surgical patients, many techniques have been proposed, based on different combinations of nutri- tional parameters obtained from anthropometry, values of hepatic secretory proteins and the determination of immune competence. In addition to being able to pre- dict surgical risk, the prognostic method should be simple and inexpensive, especially when used as a screening tool. In 1979 Seltzer et al, [6] proposed the instant nutri- tional assessment (INA), a method including only those indices derived from routine hospital laboratory tests. Mullen et al [7] developed the prognostic nutritional index (PNI), using several parameters. Although these methods were correlated with postoperative morbidity, Baker reported that clinical evaluation provided just as good prognostic information in the surgical patient [8]. Roy, who studied the predictive ability of Weight Loss t WL), PNI and INA suggested that WL alone can be used as a rapid and inexpensive method for predicting post-operative complications [9]. Recently, our group proposed a nutritional assessment (NA) method, based on serum albumin (SA), total iron binding capacity (TIBC) and weight loss; patients with at least 1 altered parameter were classified as being at high surgical risk IlO]. The aim of this work was to evaluate the prognostic ability of PNI, INA, NA and WL in predicting post- operative septic complications in a series of patients, who underwent major gastro-intestinal surgical pro- cedures for malignant disease. PATIENTS AND METHODS Between April 1986 and September 1987 the nutri- tional status, on admission was assessed in 112 con- secutive patients with gastro-intestinal cancer. Eighteen patients were subsequently excluded from the study: 14 patients because they did not undergo major surgery, 4 patients because of obvious intra-operative contamination or technical problems due to the use of stapling devices. Table 1 shows the diagnosis and oper- ative procedure in the remaining 94 patients, 43 females and 51 males, mean age 61.92 f 12.53 (SD) yr (range 33-79). Table 2 shows the 3 multiparameter methods for the identification of high-risk surgical patients, considered in this paper. Pre-operative evaluations were performed as follows. Triceps skinfold thickness (TSF) was measured at the midpoint between the olecranon and the acromion processes in the non-dominant arm, using Lange skin- fold calipers. 197