Elderly Nutritional Indicators for Geriatric Malnutrition Assessment (ENIGMA): Development and validation of a nutritional prognostic index.

BACKGROUND

Few nutritional measurement tools have been validated that predict long-term mortality risks in community-living older persons.

OBJECTIVE

To develop and validate a new nutritional prognostic index (ENIGMA) for use in geriatric outpatient or primary care.

DESIGN

We developed the ENIGMA (four questions: unable to shop, cook or feed one's self, difficulty eating due to oral problem, eat few fruits or vegetables, 3 or more drugs a day, and four blood indicators: low albumin, hemoglobin, total cholesterol and lymphocyte count), and evaluated its predictive validity for 10 years mortality outcome in a development cohort (N = 1550) of community-living older persons, in comparison with the Geriatric Nutritional Risk Index (GNRI), Mini Nutritional Assessment (MNA) and ESPEN Malnutrition (ESPEN-M), and calibrated it externally in a validation cohort (N = 924).

RESULTS

In the development cohort, ENIGMA component indicators and summary risk score (0-10) were independently associated with significantly increased mortality hazard ratio (HR), adjusted for age, sex, chronic diseases, comorbidity and inflammatory status. Increasing risk categories predicted increasing adjusted HRs (95% CI); low (0-1): reference, moderate (2-3): 1.48 (1.10-2.00), high (4-5): 2.32 (1.52-3.55), very high (6+): 4.97 (2.52-9.77). ENIGMA showed better discriminatory accuracy (C = 0.67, 95% CI = 0.63-0.71) than MNA (C = 0.59, 95% CI = 0.55-0.63), GNRI (C = 0.57, 95% CI = 0.52-0.61), and ESPEN-M (C = 0.52, 95% CI = 0.48-0.56). The predictive accuracy and utility of ENIGMA was supported in the validation cohort (C = 0.68, 95% CI = 0.62-0.74); calibration-at-large, a = 0.00007, p = 0.187; calibration slope = 0.997, 95% CI, 0.997-0.998).

CONCLUSION

The ENIGMA is a validated nutritional prognostic tool that strongly predicts long-term mortality risks and is recommended for use in geriatric outpatient and primary care settings.