Background and aims: No gold standard exists for nutritional screening/assessment. This cross-sectional study aimed to collect/use a comprehensive set of clinical, anthropometric, functional data, explore interrelations, and derive a feasible/sensitive/specific method to assess nutritional risk and status in hospital practice.
Patients and methods: 100 surgical patients were evaluated, 49M:51F, 55 +/- 18.9 (18-88) years. Nutritional risk assessment: Kondrup's Nutritional Risk Assessment, BAPEN's Malnutrition Screening Tool, Nutrition Screening Initiative, Admission Nutritional Screening Tool. Nutritional status: anthropometry categorised by Body Mass Index and McWhirter & Pennington criteria, recent weight loss > 10%, dynamometry, Subjective Global Assessment.
Results: There was a strong agreement between all nutritional risk (k = 0.69-0.89, p < 0.05) and between all nutritional assessment methods (k = 0.51-0.88, p < or = 0.05) except for dynamometry. Weight loss > 10% was the only method that agreed with all tools (k = 0.86-0.94, p < or = 0.05), and was thereafter used as the standard. Kondrup's Nutritional Risk Assessment and Admission Nutritional Screening Tool were unspecific but highly sensitive (> or = 95%). Subjective Global Assessment was highly sensitive (100%) and specific (69%), and was the only method with a significant Youden value (0.7).
Conclusions: Kondrup's Nutritional Risk Assessment and Admission Nutritional Screening Tool emerged as sensitive screening methods; the former is simpler to use, Kondrup's Nutritional Risk Assessment has been devised to direct nutritional intervention. Recent unintentional weight loss > 10% is a simple method whereas Subjective Global Assessment identified high-risk/undernourished patients.