Undernourished surgical patients suffer more frequent postsurgical complications (CP), which in addition to increasing health risks involve added expenses. The objective of this study is to evaluate the predictive value for the presence of CP, of some simple, economic malnutrition (MN) parameters that in a previous study showed an acceptable degree of correlation with the CP, and to estimate the economic savings derived from preventing, complications through diagnosis and early treatment of malnutrition. The sample consisted of all patients that over a one year period occupied a randomly selected 4-bed hospital surgical ward of a Comunidad Valenciana hospital. The clinical evolution and analytical and anthropometrical nutritional parameters were evaluated both at admission and upon discharge from hospital. The increase in expenses, due to presumably preventable CP secondary to MN is calculated from the number of Patient Attention Units generated (UMA's). Nutritional parameters showing a statistically significant correlation with CP are: serum albumin (AS); total lymphocytes (LT); tricipital fat skinfold (PGT); medial arm circumference (CB) and the ideal weight percentage (% PI). For all of them the confidence intervals of relative (RR) and attributable (RA) risks of CP when the patients show abnormal values in one or several of the nutritional parameters mentioned were calculated. The sensitivity, specificity and efficacy of each diagnostic test were estimated, as well as the positive (VP+) and negative (VP-) predictive values or conditional probabilities of suffering the disease in accordance with the positive or negative result of the diagnostic test. The length of stay in the hospital (EH) is directly related to MN, and this disorder increases remarkably the risk of suffering CP. AS and LT are the best values for predicting CP. When both are normal there is a 99% probability of not suffering CP and when both are low there is a 90% probability of developing CP, in particular nosocomial pneumonia (NEU). The EH resulting from the NEU suffered by the entire group of sample patients was 132 days, of which 112 correspond to patients having a low AS value upon admittance. Since the efficacy of AS determination is 93.5%, 105.7 days of EH secondary to predictable NEU could have been avoided, and this represents more than 8% of total yearly expenses of the General Surgical Unit studied. As and LT determination is simple, cheap and automated and allows detection of patients with a high risk of suffering MH. When surgery is elective it is possible to improve the patient's nutritional status prior to the operation and thus avoid many CP. This parameters should therefore be measured before undertaking surgery.