Nutritional risk factors for postoperative complications in Brazilian elderly patients undergoing major elective surgery

Objective In this prospective study, we assessed nutritional and immunologic risk factors for infectious complications and deaths related to infection in elderly patients undergoing major elective surgery. Methods Seventy patients 60 y or older were enrolled in this study. The preoperative variables analyzed were body mass index, body mass index knee height, triceps skinfold, subscapular skinfold, mid-arm muscle circumference, mid-arm muscle area, albumin, transferrin, prealbumin, and retinol-binding protein levels, immunoglobulins G, A, and M, C3, and C4 levels, total lymphocyte counts, and the occurrence of delayed hypersensitivity reactions (multitest). Results Abnormally low levels of prealbumin (P = 0.004), retinol-binding protein (P = 0.05), and transferrin (P = 0.04) were related to infectious complications. Prealbumin levels (P = 0.02) and lymphocyte counts below 1500 cells/mm3 (P = 0.04) were associated with mortality secondary to infection. Univariate regression analysis showed that levels of prealbumin (P = 0.02, odds RATIO = 13.3, 95% confidence LIMITS = 1.6, 110.9), retinol-binding protein (P = 0.03, odds RATIO = 4.8, 95% confidence LIMITS = 1.2, 19.3), and transferrin (P = 0.03; odds RATIO = 4.2, 95% confidence LIMIT = 1.2, 15.6) were associated with infectious complications. Multivariate analysis associated only prealbumin levels with infectious complications (P = 0.02, odds RATIO = 13.3, 95% confidence LIMIT = 1.6, 110.9). Regression analysis provided no conclusion regarding mortality because of the small number of deaths recorded. Conclusions In patients with a good cardiac index (Goldman I and II) who underwent major elective surgery, prealbumin protein, retinol-binding protein, and transferrin levels below normal values represented a significant risk for postoperative infectious complications. Lymphocyte counts lower than 1500/m3 and abnormal prealbumin values were associated with postoperative mortality secondary to infection. The anthropometric variables evaluated did not predict postoperative infectious complications and mortality.