CEREDA/c/pppp:04 LORD_c

Objectives: To investigate if the use of estimeted height (EH) by currently available prediction formulas might affect the screening and outcome prediction attitudes of both the Mini Nutritional Assessment (MNA) and its short-form version (MNA-SF). Design: A 6-month observational study. Setting: Two long-term cares of the province of Como. Participants: 266 resident elderly (102 men, 164 women; mean age±SD: 80.4±8.6 years). Measurements: Subjects were studied by anthropometry (weight, standing height, knee-height, arm and calf circumferences, triceps skinfold) and biochemistry (albumin and prealbumin). Nutritional status was assessed using both MNA and MNA-SF. At 6 months, major outcome were: death, infections and bedsores. Results: In overall population, prediction formulas significantly underestimated real height. The bias by Italianspecific equation was higher than that by nationally-representative formulas for white Americans. The use of EHs produced significant differences in body mass index (BMI) but these did not affect nutritional status scoring by MNA and MNA-SF (r≥0.99, p<0.0001). Cohen’s kappa statistic also showed an almost perfect concordance (κ>0.9). Moreover, similar degrees of correlation were found between nutritional parameters and both MNA and MNA-SF scores by BMI from SH and EHs. After 6 months, major complications occurred in twenty-eight patients (11.6%). The use EHs did not affect the distribution of events among MNA and MNA-SF nutritional classes. Conclusion: In Italian elderly, height prediction by nationally representative equations for white Americans should be preferred to that by ethnic-specifc formula. However, the use of both models does not seem to affect nutritional screening and outcome prediction by MNA and MNA-SF.