Assessment of Energy Consumption Patterns and Validation of Predictive Equations among ICU Patients at Muhimbili National Hospital in Dar es Salaam, Tanzania

Abstract

Background: Nutritional support in critically-ill patients is one of the most important parameters guarding the prognosis and influencing morbidity and mortality in these patients, owing to that fact, accurate measurement of the resting energy expenditure using Indirect calorimetry is recommended by guidelines as a gold standard. But due to lack of resources and other technicalities predictive equations are conveniently used as surrogates. Aims: This study was intended to examine the extent of malnutrition, poor nutrition support practices and to validate the common used predictive equations in critically-ill patients in our setting. Methodology: A hospital-based descriptive cross-sectional study was conducted on consecutively sampled 110 mechanically ventilated ICU patients at Muhimbili National Hospital Mloganzila. Anthropometric measurements, duration of stay in ICU, Temperature and Minute Volumes were recorded so as to estimate resting energy expenditure, REE from different predictive equations. Using Indirect Calorimetry Module patient’s REE was measured and recorded then a statistical data analyzed using SPSS software version 23. Results: The prevalence of poor nutritional support was 69%; underfeeding and overfeeding were observed in 41.8% and 27.3% of all participants respectively. Prevalence of malnutrition was 51.8%; underweight and overweight were found to be in 29.1% and 22.7% of all participants respectively. The accuracy of predictive equation in ±10% difference was 30%, 45.5%, 46.4% and 68.2% in HB, MSJ, ESPEN and PENN respectively. Conclusion and Recommendation: Malnutrition and Poor Nutritional support are common problems in ICU. Predictive equations have poor accuracy and validity in comparison to indirect calorimetry. Penn State Equation was the most accurate and with the highest agreement with IC.