Energy expenditure (EE) in critically ill patients is highly variable depending on the initial severity of the disease and treatments. Clinicians need to measure EE by indirect calorimetry (IC) to optimize nutritional support. IC devices available on the market have insufficient accuracy for clinical and research use. A new IC is being developed to meet these needs.

Objectives: Validation of a reliable and easy-to-use IC device.

Primary objective:

• Accuracy of EE measurements (kcal/d measured over 30 min) of the new and the reference IC devices in intubated ICU patients on mechanical ventilation

Secondary objectives:

• Accuracy of EE measurements in ICU patients of spontaneous breathing
• Accuracy of EE measurement during ~30 min. vs. the sum of 6 periods of 60 min/12 hrs
• Accuracy of EE measurements in ICU patients on non-invasive mechanical ventilation

Design: Prospective, controlled, observational multicenter study.

Subjects: 182 ICU patients.

Energy expenditure (EE) in critically ill patients is highly variable depending on the initial severity of the disease and treatments. Clinicians need to measure EE by indirect calorimetry (IC) to optimize nutritional support. IC devices available on the market have insufficient accuracy for clinical and research use. A new IC is being developed to meet these needs.

Objectives: Validation of a reliable and easy-to-use IC device.

Primary objective:

• Accuracy of EE measurements (kcal/d measured over 30 min) of the new and the reference IC devices in intubated ICU patients on mechanical ventilation

Secondary objectives:

• Accuracy of EE measurements in ICU patients of spontaneous breathing
• Accuracy of EE measurement during ~30 min. vs. the sum of 6 periods of 60 min/12 hrs
• Accuracy of EE measurements in ICU patients on non-invasive mechanical ventilation Design: Prospective, controlled, observational multicenter study. Subjects: 182 ICU patients. Sample size calculation is based on the estimated accuracy of the new IC device compared to the Deltatrac, when performed simultaneously on one subject. In a previous study, we found a difference of EE between the Deltatrac and the Quark RMR of 24±220 kcal, thus a coefficient of variation (SD/mean) of 8.8 %. We hypothesized a difference of 25±50 kcal when comparing the Deltatrac with the new IC device, thus a coefficient of variation of 2%. This reduction of 6.5% of the coefficient variation, with a power of 90% and a significance level of 0.05 was used to calculate a required sample size of 91 subjects. A drop out rate of 50% is anticipated, leading to 182 patients to be recruited to allow an analysis based on 91 patients with optimal measurements.

Inclusion criteria:

• Adult ICU patients with spontaneous breathing, or on invasive or non-invasive mechanical ventilation.

Exclusion criteria:

• FiO2 > 60%
• Kidney failure requiring Continous veno-venous hemofiltration (CVVH)
• Conditions promoting air leaks: Positive end-expiratory pressure (PEEP) >10, Peak ventilatory pressure > 30
• Chest tube (except if only for liquid drainage = no bubbling)
• Unstable hemodynamic conditions requiring changes in vasoactive drugs either 1 hour before and/or during the IC measurement (i.e. all conditions resulting in unstable pH during IC measurements).
• Change in body temperature >0.5°C either 1 hr before and/or during the IC measurement
• Physical agitation or curarization either 1 hour before and/or during the IC measurement

Outcomes:

Primary outcome:

• EE in intubated ICU patients on mechanical ventilation

Secondary outcomes:

• EE in ICU patients of spontaneous breathing
• EE during ~30 minutes vs. the sum of 6 periods of 60 minutes/12 hours
• EE measurements when using different types of ventilators
• Time to start EE measurement (calibration, warming up)
• Duration for IC disinfection New Calorimeter - Protocol - November 11. 2013 Page 3 sur 5
• Limitations of EE measurements in patients on non-invasive mechanical ventilation

Indirect calorimetry devices and methodology:

• New IC device (Cosmed, Italy) and Deltatrac II (Datex, Finland).
• Briefly, patients are laying down in a supine position on a bed during ongoing measure of ~30 min (see Figure 1) by connecting the calorimeter to the endotracheal tube (invasive mechanical ventilation) or to the facemask (non-invasive mechanical ventilation), or by using a transparent canopy in plexiglas to cover the head while room-air flows through and is mixed with the expirate (spontaneous breathing). This mixture of patient breath and room air is finally collected and analyzed by the IC device for O2 and CO2 concentrations, and used to calculate EE. The first 2 min of data is eliminated as acclimatization artefact. From the remaining ~28 min, a segment of 5 consecutive 1-min measures with 5% coefficient of variation in VO2 and VCO2 is considered as steady state.

Barometric pressure is recorded. EE is calculated using the modified Weir equation.

• Calibration of the gas analyzers is carried out before /after each measurement according to the manufacture instructions, including ethanol burning for the Deltatrac II.
• EE measurements is performed simultaneously by the 2 IC devices.

Data collected:

• Sex, Age (yr), height (±2 cm), dry weight (±1kg)
• Heart rate, Body temperature
• Admission diagnosis, Major treatments
• SAPS II and SOFA on ICU admission day; SOFA on the study day
• Ventilator type (name) and ventilation mode (Minute ventilation, FiO2)
• Inspiratory and expiratory O2, CO2, air flow (ml/min) measured by the IC

Statistics and data analysis:

The Coordinating center analyses the data transferred by the Investigation centers

• All data are reported as means±SD. The parameters (see above) of the two IC devices is compared by Bland and Altman plots
• Reliability and adequacy (sensitivity, specificity, positive and negative predictive values) between the IC devices, is tested using ROC curves with kappa coefficient (reliability coefficient). For coefficient of variation, ANOVA is used when applicable. Results are adjusted using Greenhouse-Geisser correction in case of non-sphericity. Post-hoc tests using Bonferroni correction compared the results between two IC devices. The prediction of EE differences between devices are analyzed by multivariate regression. As 91 are expected to complete the entire protocol, 9 factors are introduced in the model (Peduzzi P, J Clin Epidemiol 1996, 49, 1373-9), i.e. ventilator type, ventilator mode, minute ventilation, FiO2, BMI, SOFA, heart rate, body temperature, medical/surgical category.

Statistical analysis significance is set at p<0.05.

For the subgroup of patients with prolonged EE measurements or on non-invasive ventilation:

Univariate and multivariate analysis are used to analyze the factors associated with observed variations.

Value of the study for clinical practice practice and research:

An accurate and reliable IC for nutrition specialists will allow to optimize energy prescription, a factor shown to improve the clinical outcome. An easy-to-use and affordable (10'000 $) IC will reduce both the burden on human resources and costs.

Study originality:

• Development of an IC device meeting technical characteristics defined by physicians (bottom-up strategy)
• Demonstration of the clinical relevance of short EE measurements, practically convenient and less human-resources consuming, versus longer EE measurements New Calorimeter - Protocol - November 11. 2013 Page 4 sur 5
• Exploratory comparative tests, whenever possible, to evaluate IC in a subgroup of patients with non-invasive ventilation New Calorimeter - Protocol - November 11. 2013 Page 5 sur 5

Agenda of actions:

• The IC prototype will be made available the 18th of May 2014.
• Recruitment of subjects will begin in May-June 2014 and will continue for 1 year, or until sufficient subjects are included in the study.

Inclusion Criteria:

• Adult ICU patients with spontaneous breathing, or on invasive or non-invasive mechanical ventilation.

Exclusion Criteria:

• FiO2 > 60%

  • Kidney failure requiring Continous veno-venous hemofiltration (CVVH)
  • Conditions promoting air leaks: Positive end-expiratory pressure (PEEP) >10, Peak ventilatory pressure > 30
  • Chest tube (except if only for liquid drainage = no bubbling)
  • Unstable hemodynamic conditions requiring changes in vasoactive drugs either 1 hour before and/or during the IC measurement (i.e. all conditions resulting in unstable pH during IC measurements).
  • Change in body temperature >0.5°C either 1 hr before and/or during the IC measurement
  • Physical agitation or curarization either 1 hour before and/or during the IC measurement

All adult ICU patients with spontaneous breathing, or on invasive or non-invasive mechanical ventilation admitted into the intesive care unit