Role of nutritional indices in predicting outcomes of vascular surgery

Background: Malnutrition is frequent among vascular surgery patients, given their age, chronic comorbidities, and poor functional status, and it is believed to increase their operative risk. We aimed to assess the combined use of recent significant weight loss (>10% body mass) and serum albumin levels as a nutritional status index to predict outcomes. Methods: We analyzed vascular surgery data from the American College of Surgeons National Surgical Quality Improvement Program database (2005-2012; N = 238,082) to compare operative death (in-hospital and 30-day operative death) across eight nutritional status groups based on weight loss (yes/no) and albumin category: very low albumin level (VL-Alb; <2.50 g/dL), low albumin level (L-Alb; 2.50-3.39 g/dL), normal albumin level (N-Alb; 3.40-4.39 g/dL), and high albumin level (H-Alb; 4.40-5.40 g/dL). Risk-adjusted odds ratios (AOR) with 95% confidence intervals were estimated by multivariable logistic regression (N-Alb [no weight loss], reference). Results: The study population included 113,936 patients for whom albumin level was available (age, 67 ± 13 years; 60.2% male). Operative death was documented in 5160 (4.53%) patients. The eight-category nutritional status was more predictive of operative death than age alone (C statistic, 0.74 vs 0.63). A high discrimination multivariable model for operative death was derived (C statistic, 0.851). Low albumin level was associated with increased death that worsened in case of weight loss: VL-Alb + WL, AOR = 3.83 (3.03-4.83); VL-Alb, AOR = 3.36 (3.06-3.69); L-Alb + WL, AOR = 2.46 (1.98-3.05); and L-Alb, AOR = 1.99 (1.84-2.15). Weight loss was associated with increased death even if albumin level was normal: N-Alb + WL, AOR = 1.77 (1.34-2.35); and H-Alb + WL, AOR = 1.91 (0.69-5.31). H-Alb was protective (AOR = 0.65 [0.55-0.76]). Conclusions: Nutritional status predicts outcomes of vascular surgery. Serum albumin level and weight loss should be incorporated in patients' risk stratification.