The Berlin-Buch respiration chamber for energy expenditure measurements

Design and layout

The respiration chamber at the Experimental and Clinical Research Center in Berlin-Buch, Germany was established in 2010. This was an interdisciplinary team effort based on over 30 years of scientific experience with indirect calorimetry (Steiniger 1984, 1985; Steiniger et al. 1987; Boschmann et al. 2007). Technical expertise was provided by the main contractor Linde GmbH (Gases Division, Berlin, Germany) and by subcontractors for dry construction, ventilation (LORMS Service AG, Ahrensfelde, Germany), measurement and control technology (digitech gmbh, Ahrensfelde, Germany) and for sensors and gas analysis (HTK Hamburg GmbH, Germany).

The chamber (length 2.5 m, width 2.0 m, height 2.2 m, total volume 11,000 L, net volume 9900 L) is a room-in-room dry construction. Thermal insulation and ventilation system are integrated between the inner and outer chamber walls. To prevent air leaks, the inner walls are sealed with a special paint (CreaGlas 2 K-PU-Finish 3471, Brillux GmbH & Co. KG, Münster, Germany). High precision sensors measure oxygen, carbon dioxide, airflow, and climatic conditions. An individual and flexible self-programmed software package based on Microsoft Visual Studio controls measurements, collects, and processes all data required to assess changes in EE and macronutrient oxidation. Raw data are transferred into several MS Excel files for further analysis.

The chamber has a great airtight window and a glass door allowing subjects to look outside (Fig. 1A). Both window and door can be covered with blinds from inside the chamber for privacy. Technical equipment is located outside the chamber to create a non-stressful environment. The chamber is equipped with air conditioning, a comfortable chair with a footrest, a table, TV rack, TV/DVD set, a bicycle ergometer, and a camping toilet (Fig. 1B). Bicycle ergometer and chair can be replaced by a bed for sleeping EE measurements. Therefore, the set-up allows measuring EE from 30 min up to 24 h.

Fig. 1figure 1

Photograph (A) and layout (B) of the pull-type, open-circuit respiration chamber in Berlin, Germany. 1, table; 2, television set; 3, cameras; 4, Passive Infrared Sensors; 5, chair with footrest; 6, bicycle ergometer; 7, toilet; 8, air lock; 9, air conditioning

Close supervision of the subject is facilitated by two mini colour cameras, one pointing to the armchair, one to the bicycle ergometer. Of note, there are no recordings at any time. If necessary, subject and investigator can communicate via an intercom. Camera supervision can be discontinued if requested by the subject.

The chamber is equipped with three passive infrared sensors (PIRS), one in front and two above the subject. PIRS react to heat emanating from moving bodies, i.e. humans and animals. If the module recognizes a movement, it produces a digital signal. The signal is registered in volt, thus allowing for semi-quantitative evaluation of activity.

The quite air conditioning (Sanyo SAP-FDRV96EH, 3 kW, 300–500 m3/h, 22–30 dB) can be adjusted by the subject from within or by the investigator from outside.

Flow rate and sensors

Fresh air is pulled into the chamber at four entry points near the ceiling, is mixed by the air conditioning, and exits at four points in the middle of the opposite wall (pull calorimeter). This air flow is facilitated by a precision pump (ORPU V01Y, Pumpenfabrik GmbH, Germany) and can be adjusted to a flow rate between 100 and 200 L/min. Flow is measured by a thermal mass flow controller (HTK Hamburg, Germany). The majority of measurements is done with a flow rate of 120 L/min. The negative pressure registered by the chamber at this flow rate is about 6 kPa.

Gas analysis is controlled by a Process Control Unit (PCU10-O2/CO2-S, HTK Hamburg, Germany). The unit contains two paramagnetic O2 sensors (range 20.000–21.000 Vol%, deficit 0.000–1.000 Vol%, HUMMINGBIRD, Servomex, UK) and two NIDR infrared CO2 sensors (0.000–1.000 Vol%, Dynament Ltd, UK), one for incoming and one for outgoing air samples. Resolution of all four sensors is better than 0.001 Vol%. Two diaphragm pumps draw aliquot samples (80 mL/min) of incoming and outgoing air to the respective gas analysis system.

Every second, flow rate, barometric pressure, temperature, relative humidity, and PIRS measurements are digitalized (DT9813-10 V, 12 bit, Data Translation Inc.) and transferred to the computer (RS 232, USB 2.0). The interface was programmed using the DT-Open Layers for NET Class Library (Data Translation Inc.). Further analysis is done with 1 min averages of these data with no further signal processing techniques.

Calibration

Before each measurement, O2 and CO2 sensors are calibrated simultaneously with three calibration gases: 20.0% O2 in N2, 21.0 O2 in N2 and 0.8% CO2 in N2. Certified accuracy is ± 0.2% for O2 and ± 0.02% for CO2 (Linde GmbH, gases division, Berlin, Germany). Gases enter the gas analysis system at a flow rate of 80 mL/min. The software triggers a 360 s calibration routine. Valves that open and close to allow calibration gas flow switch at 0 s (O2 zero, 20%), 120 s (O2 span, 21% and CO2 zero, 0%), 240 s (CO2 span, 0.803%), and 360 s (end). Calibration analysis is based on data collection during concentration plateaus after 120 s for O2 sensors and after 100 s for CO2 sensors. All calibration data are saved in MS Excel files to monitor precision and stability of all gas sensors over time.

Validation

To verify the accuracy of \(\dot}_\), \(\dot}_}\), EE, and respiratory exchange ratio (RER) measurements, the calorimetric system is validated every 2 weeks by 2 h acetone burning tests. For this, a wick lamp filled with acetone (purity > 99%) is placed on a digital scale and ignited. The scale (maximal load 1000 g, resolution 0.01 g, KERN PCB 1000–2, Kern & Sohn GmbH, Germany) is connected to the computer by a RS 232 interface in order to register the amount of burned acetone.

Theoretically, burning of 1 g acetone (corrected for acetone content of 99.95% and evaporation of 0.0066 g/min (own data)) consumes 1.545 L O2 and produces 1.150 L CO2. Consequently, RER (\(\dot}_}\)/\(\dot}_\)) and EE should be 0.746 and 30.42 kJ, respectively. Calculated values are divided by measured values during the validation to receive validation factors (calculated/measured). The factors for \(\dot}_\) and \(\dot}_}\) are used to correct subject measurements within the respective validation interval (usually 2 weeks).

In addition to regular 2 h validations, we do 20 h measurements that consist of three phases: an acetone burning phase for validation purposes (~ 8 h), a washout phase to determine chamber volume (~ 7 h), and a comparison phase to evaluate sensor stability over time (~ 5 h).

Calculations and algorithms

\(\dot}_\) and \(\dot}_}\) are calculated from gas volumes and climatic data with fixed averaging windows of 5, 10, or 15 min using the equations by Brown et al. (1984). From these volumes, EE is calculated using glucose, fatty acids, and amino acids as reference systems. Stoichiometric factors for carbohydrate oxidation are derived from glucose (C6H12O6) and for fat oxidation from palmitoyl-stearoyl-oleoyl-glycerol (C22H104O6) according to Ferrannini et al. (Ferrannini 1988). For protein, we use the nine amino acids most relevant for energy metabolism (ALA, ASP, ASN, GLU, GLN, LEU, ILE, ARG, LYS; combined formula C5.0H10.6O2.7N1.7). We presume that 100% of nitrogen liberated in protein metabolism is excreted as urea via the urine.

(1)

Glucose (g) = − 3.200 \(\dot}_\)(L) + 4.541 \(\dot}_}\) (L) − 2.688 Nex (g); (15.65 kJ/g)

(2)

Palmitoyl-stearoyl-oleoyl-glycerol (g) = 1.669 \(\dot}_\) (L) − 1.669 \(\dot}_}\) (L) − 1.414 Nex (g); (39.75 kJ/g)

(3)

Amino acids (g) = 5.848 Nex (g); (16.14 kJ/g)

(4)

EE (kJ/min) = 16.28 \(\dot}_\) (L) + 4.70 \(\dot}_}\) (L) − 3.88 Nex (g)

Under resting conditions and diets low in protein, a simplified formula can be used since nitrogen excretion (Nex) is nearly constant at 5.33 mg/min (Steiniger et al. 2009).

(5)

EE (kJ/min) = 16.07 \(\dot}_\) (L/min) + 4.89 \(\dot}_}\) (L/min) − 0.035

Statistics

Statistical analyses were performed with GraphPad Prism (Version 9.2.0). Data are given as mean ± SD. Gas sensor variablilty was evaluted by simple linear regression. ROUT method was used to detect outliers with a desired maximum false discovery rate of 1%.

留言 (0)

沒有登入
gif