PD015 - PREDICTIVE EQUATIONS MISS THE MARK: ENERGY PRESCRIPTION INACCURACY AND COST IMPLICATIONS OF INDIRECT CALORIMETRY IN CRITICALLY ILL PATIENTS

PD015

PREDICTIVE EQUATIONS MISS THE MARK: ENERGY PRESCRIPTION INACCURACY AND COST IMPLICATIONS OF INDIRECT CALORIMETRY IN CRITICALLY ILL PATIENTS

H. Cunha^1,*, V. Ludtke¹, G. Freitas¹, C. D´Avila¹, N. Gallagher¹

¹Hospital Vitória / Samaritano , Rio de Janeiro, Brazil

Rationale: Predictive equations (PE) are routinely used for energy prescription in ICU patients when indirect calorimetry (IC) is unavailable. The clinical and economic consequences of PE inaccuracy remain poorly characterized.

Methods: Cross-sectional single-day observational study of 26 mechanically ventilated adults on enteral nutrition in a single ICU. Energy expenditure was measured by IC (gold standard) and estimated by PE (kcal/kg/day). Patients were classified as underfed (PE < IC), overfed (PE > IC), or concordant (error ≤10%). Daily enteral nutrition costs were calculated using institutional prices.

Results: Regarding EN product costs, correcting underfeeding to IC-measured targets required a mean additional expenditure of R$37.09±37.21/patient/day, while correcting overfeeding generated mean savings of R$16.04±16.92/day. Net EN product cost differential: +R$18.70/patient/day. IC operational costs were not captured in this model.

Conclusion: PE failed to achieve energy targets in 65% of critically ill patients, predominantly through underfeeding (mean deficit 383 kcal/day). IC-guided correction of underfeeding carries a mean EN product cost increase of R$37/day — reflecting the price of nutritional adequacy, not waste. This analysis was limited to EN product costs; IC operational costs (equipment, disposables, technician time) were not modeled, and full cost-effectiveness requires incorporating these. This cross-sectional single-day design also does not capture prescription variability over the ICU stay. These findings establish the EN cost differential as foundation for a complete economic model of IC utilization.

Disclosure of Interest: None declared