Validity of the Lumen® hand-held metabolic device to measure fuel utilization
in healthy young adults
Kent A. Lorenz1*, Shlomo Yeshurun2, Richard Aziz1, Julissa Ortiz-Delatorre1, James R. Bagley1, Merav
Mor2*, and Marialice Kern1
1Exercise Physiology Laboratory, San Francisco State University, Department of Kinesiology, San
Francisco, CA USA.
2Metaflow Ltd., Tel Aviv, Israel
Kent A. Lorenz
Keywords: Resting Metabolic Rate, Lumen®, ParvoMedics TrueOne® 2400, Validation, Respiratory
Exchange Ratio, Metabolism, Fuel Utilization, Indirect Calorimetry
Background: Metabolic carts measure the carbon dioxide produced and oxygen consumed from the
breath in order to assess metabolic fuel usage (carbohydrates vs. fats). However, these systems are
expensive, time-consuming, and only available in the clinic. A small hand-held device capable of measuring
metabolic fuel via CO2 from exhaled air has been developed. Objective: To evaluate the validity of a novel
hand-held device (Lumen®) for measuring metabolic fuel utilization in healthy young adults. Methods:
Metabolic fuel usage was assessed in healthy participants (n = 33; age: 23.1 ± 3.9 y) via respiratory
exchange ratio (RER) values from the “gold-standard” metabolic cart as well as %CO2 from the Lumen
device. Measurements were performed at rest in two conditions, fasting, and after consuming 150 grams
of glucose in order to determine changes in metabolic fuel. Reduced major axis regression was performed
as well as Bland-Altman plots and linear regressions to test for agreement between RER and Lumen %CO2.
Results: Both RER and Lumen %CO2 significantly increased after glucose intake compared with fasting
conditions (P < .0001). Regression analyses and Bland-Altman plots revealed an agreement between the
two measurements with a systematic bias resulting from the nature of the different units. Conclusions:
This study shows the validity of Lumen® to estimate metabolic fuel utilization in a comparable manner
with the “gold-standard” metabolic cart, providing the ability for real-time metabolic information for users
under any circumstances.
Respiratory quotient (RQ) is the ratio of carbon dioxide produced to oxygen consumed (VCO2/VO2)
measured directly at the cellular level to estimate metabolic fuel utilization. However, this method
requires the insertion of a catheter into the vein and artery for a blood sample or taking a tissue sample,
which makes the RQ measurement invasive and infeasible outside of laboratory setting . With the use
of a metabolic cart, an indirect measure of RQ can be made by indirect calorimetry through respiratory
gas exchange of O2 and CO2. This is the respiratory exchange ratio (RER), which is currently the preferred
method for determining metabolic fuel utilization. Unlike RQ, RER measures the carbon dioxide produced
(VCO2) and oxygen consumed (VO2) from exhaled air [1,2]. Both RQ and RER indicate the relative
contribution of carbohydrate and lipid to energy expenditure . Though RER measurement is not
invasive, this method is time-consuming (up to 40 minutes) and is only available in test laboratory setting
and it requires technical and physiological expertise for handling the metabolic cart and interpretation of
the metabolic data obtained.
Metaflow Ltd. developed Lumen®, a novel metabolic fuel utilization breathalyzer which is a personalized
hand-held device that provides an individuals’ metabolic state in real-time, by measuring CO2 from
exhaled breath (Figure 1). The device indirectly measures metabolic fuel usage via a CO2 sensor and a flow
sensor, to determine the rate of CO2 production from a single breath maneuver. The CO2 concentration in
the exhaled volume of air is determined from a specific breathing maneuver with a breath hold of 10
seconds. This concept is based on the fact that oxygen consumption is stable under resting conditions ;
thus, a change in the metabolic fuel use will generally be represented by changes in CO2 production. For
carbohydrate oxidation more carbon dioxide is produced, relative to the consumption of oxygen. For fat
oxidation, less carbon dioxide is produced. Thus, from the resulting RER, for a resting state ranging
between 0.7 and 1, the contribution of lipid or carbohydrate usage can be determined. The breath
maneuver enables quantifying the changes in CO2 production, allowing the user to estimate their
metabolic state . The use of a smartphone application enables the user to track metabolic status
outside of physiologic test laboratories.
Internal studies (unpublished data) showed the potential of Lumen to accurately measure the metabolic
fuel usage in response to diet and exercise, comparable to the “gold standard” metabolic cart. In this
study, we aim to evaluate agreement between Lumen measurement and metabolic cart in healthy
participants before and after glucose ingestion under stable resting conditions.
Figure 1. A schematic representation of the Lumen® device and application.
Fifty-four healthy volunteers reported to the Exercise Physiology Laboratory from the Department of
Kinesiology at San Francisco State University to participate in this study. To be included for the study,
participants must have been between the ages of 18-45 years old; with a BMI less than 30 kg/m2;and not
doing any high intensity aerobic exercise training over 3 times per week and without any known
cardiovascular-, pulmonary-, and/or metabolic diseases. The study was approved by the University’s
Institutional Review Board for Human Subjects, and written informed consent was obtained from each
participant before testing.
Participants were recruited and their height and weight measured using a stadiometer and Seca scale
(Seca, Hamburg, Germany). If they met the BMI criteria, they were provided their own Lumen device
which was labeled with their unique identification number. The Lumen device was paired and
synchronized to the participant smartphone together with the Lumen application. Participants practiced
the Lumen breathing technique while supervised and took the device home for further familiarization
period in order to show proficiency with the device and application. They were instructed to perform
Lumen metabolic measurements for at least 30 sessions, with each session consisting of 3 breath
maneuvers, and to take 3 sessions at different time points each day. After the minimum amount of home
breath sessions were collected, participants were scheduled for the study laboratory measurement day.
All participants came to the test laboratory between 07:00 a.m. and 11:00 a.m. after a 12-h fast and had
abstained from any form of physical activity (other than walking).
On the laboratory testing day, blood glucose samples were taken by sterile finger prick blood sample and
measured by a glucometer (OneTouch, LifeScan Inc. Milpitas, CA). For the indirect calorimetry
measurement, the participant had to lay down in supine position on a padded examination table, where
a rigid clear plastic canopy with a comfortable, flexible seal was placed over the head and upper part of
the torso. Once the metabolic cart measurement was ended, the participant was seated in a comfortable
chair. After five minutes of rest, they were asked to perform two Lumen breath sessions (5-minute break
between each session). A valid Lumen session measurement for the evaluation with a difference of less
than 0.2 %CO2 between the breaths in a session. The first Lumen session immediately after the metabolic
cart measurement was used for data analysis. In case with an invalid first session, the second session was
Once finished, participants were asked to drink 150 grams of a glucose solution (3 servings of 50 grams
with 20 minutes intervals between each serving). Forty-five minutes after the intake of the first drink
(corresponding to 5 minutes after finishing the last serving), their glucose levels were reassessed, and the
same assessment procedures as during the fasted state before the glucose intake were repeated.
Participants were removed from the analysis if they were unable to finish all glucose drinks.
RER was analyzed using a calibrated TrueOne® 2400 metabolic cart (ParvoMedics, Murray, UT, USA). This
system uses a paramagnetic oxygen analyzer and infrared carbon dioxide analyzer with a Hans Rudolph
heated pneumotach. The ParvoMedics system was warmed up for at least 60 minutes each day before
testing, in order to ensure accurate and stable readings. The gas analyzers and flow sensor were calibrated
as per manufacturer’s recommendations. Calibration of the analyzers performed using a high precision
gas mixture (O2, CO2, remainder N2) and calibrated and accepted with a less than 0.1% error with the
calibration gas. Flow and volume were calibrated using a calibrated 3 L syringe (Hans Rudolph, model
5530) to ≤ 1% error. The ambient temperature was kept between 22 and 26°C in the test laboratory.
Relative humidity was maintained stable at roughly 60%. Once calibration was acceptable and complete,
a ventilated hood with subject cover was placed over the participant’s head and positioned around the
upper torso area to ensure no escaping air from the hood. The participants were required to stay awake
during the measurement procedure. The hood ventilation (VE) was measured during the recording, and
CO2 and O2 concentrations were measured from it. VCO2, and VO2 parameters were calculated and taken
as 30-s averages. For this study we defined the subject steady-state metabolic measurement based on
observed variations in the VO2 and VCO2 of less than ≤ 5% CV for period of at least five consecutive
minutes. Inability to meet these criteria resulted in removal of the data from the analysis.
During the measurement day, participants took 2 sessions of 3 Lumen breaths after the metabolic cart.
The Lumen breathing maneuver consists of three phases, starting from the end of a normal expiration
(functional residual capacity). The participant takes a deep breath in through the Lumen device, followed
by a 10 second breath hold. Afterwards, the subject exhales through the Lumen device, with a steady
exhalation flow to at least the starting level of the maneuver. The Lumen smart phone application guides
the participant through each phase of the lumen maneuver. Each Lumen session was repeated after a 5-
minute pause interval. Validity of breath maneuvers was systematically evaluated by the Lumen
application. Inability to perform valid Lumen breath measures resulted in removal of the data from the
All variables were tested and visualized for normal distribution before the tests.
To evaluate the changes after glucose intake, two-tailed paired parametric t-tests were performed for
blood glucose levels, RER levels, and Lumen %CO2 before and after glucose intake.
For agreement validations, major axis regression (‘Deming’s method’) was performed in order to compare
RER of the metabolic cart and %CO2 from the Lumen device . As RER and %CO2 are in different units,
the analysis is identical to ordinary least product regression (also known as reduced major axis regression),
which is the most suitable analysis for comparison between two methods of measurement . Moreover,
Bland-Altman plot was created to demonstrate limits of agreement, together with a linear regression to
test the distribution of error to identify if any systematic bias existed between Lumen and RER measures.
In addition, a simple linear regression (ordinary least squares) was performed to determine the ability to
predict Lumen values from the gold-standard value of RER.
Statistical analyses were performed using GraphPad Prism 8 (GraphPad Software Inc., LA Jolla, CA). The
threshold for significance was set at P < .05.
From the original fifty-four participants recruited, twelve were excluded prior to laboratory testing and
nine had to be excluded during the testing day for failing to meet the inclusion criteria as detailed in the
methods section: one participant was unable to consume all glucose drinks due to nausea, three
participants did not achieve 5 mins of stable metabolic cart measurement (CV < 5% in VO2 and VCO2), and
five participants were unable to perform a valid Lumen measurement (Figure 2). Characteristics of the
final thirty-three participants presented in Table 1.
Figure 2. Consolidated Standards of Reporting Trials (CONSORT) flow diagram.
Table 1. Descriptive statistics of study’s participants.
24.0 ± 3.0
73.7 ± 10.2
171.7 ± 7.8
24.9 ± 2.5
22.3 ± 4.5
59.1 ± 6.4
160.9 ± 5.5
22.9 ± 2.6
23.1 ± 3.9
66.2 ± 11.1
166.1 ± 8.6
23.9 ± 2.7
Data are presented as mean ± SD.
Blood glucose levels increased from 90.6 ± 9.2 mg/dL to 145.2 ± 25.3 mg/dL as a result of glucose intake
(t(32) = 11.04, P < .0001; Figure 3A). RER levels increased from 0.787 ± 0.043 to 0.876 ± 0.053 in response
to glucose intake (t(32) = 10.84, P < .0001; Figure 3B). Moreover, Lumen %CO2 concentrations significantly
rose from 4.20 ± 0.4 to 4.48 ± 0.34 (t(32) = 5.978, P < .0001; Figure 3C). These analyses have confirmed the
ability of both the metabolic cart and Lumen to detect changes in metabolic fuel utilization.
Figure 3. Changes of blood glucose (A), RER (B), Lumen %CO2 (C), after glucose intake. Data are presented as mean
± SD. ****p < 0.0001. n = 33 for each state.
In order to look for agreement between RER units from the metabolic cart and %CO2 from Lumen, reduced
major axis regression was performed . It revealed a significant relationship between RER and Lumen
%CO2 (F(1,63) = 18.54, P < .0001, y = 6.111x - 0.7445, x-intercept = 0.1218; Figure 4). This analysis validated
the agreement between Lumen %CO2 and metabolic cart RER, with a systemic bias resulting from the
nature of the different units.
Figure 4. Reduced major axis regression of RER from the metabolic cart and Lumen’s %CO2 measurements for
metabolic activity. n = 33 for each state.
A Bland-Altman plot was made to discuss limit of association between RER and Lumen %CO2 . Bland-
Altman plots are used to calculate the level of agreements between two measures by studying the mean
difference between measurements and constructing limits of agreement . Bland-Altman analysis
revealed a mean difference between RER units and Lumen %CO2 of 3.505 with 95% limits of agreement
between 2.784 to 4.226 (Figure 5). Since the Lumen device always provides a measurement that is
numerically much higher than RER, Lumen measurements with smaller values should produce smaller
differences and larger Lumen values will produce larger differences. To test the bias in measurements, a
regression of difference (Lumen %CO2 – RER) scores on average ((Lumen %CO2 + RER)/2) values was
performed. Simple linear regression showed that a significant model effect of the bias (F(1,63) = 721, P <
.0001, R2 = 0.9196). As the vertical distribution of scores are tight to the line, it indicates that the errors
are consistent across a range of values.
Figure 5. Bland-Altman analysis with simple linear regression. The solid line represents the mean bias between the
Lumen %CO2 and RER. The upper and lower dashed lines represent the 95% confidence intervals (±2 SD) from the
mean bias. Linear regression line: y = 1.646*x - 0.7472. n = 33 for each state.
To determine the ability of RER to predict Lumen %CO2, ordinary least squared regression was performed
to estimate Lumen values from RER measures, with the assumption that RER is an accurate measure. With
RER as the independent variable, we used linear regression to predict Lumen %CO2, and a significant
model effect was present (F(1,63) = 18.54, P < .0001, R2 = 0.2274; Figure 6). The RER parameter estimate
indicated that for every one-unit increase in RER, a 2.914-unit increase (SE = 0.6767) in Lumen %CO2 is
expected. However, since a full unit increase in RER is not a plausible outcome, this parameter estimate
can be interpreted similarly by saying a 0.1-unit increase in RER (e.g., 0.7 to 0.8) will produce a 0.2914-
unit increase in Lumen %CO2.
Figure 6. Ordinary least squared regression of RER and Lumen %CO2. n = 33 for each state.
This study evaluated the capability of the Lumen device to assess changes in the body’s metabolic fuel
utilization, compared to gold standard indirect calorimetry metabolic cart measurement in healthy young
adults. Our results show that Lumen %CO2 levels are in agreement with RER values from the metabolic
cart, which correspond to relative changes in metabolic fuel utilization.
Both Lumen %CO2 and metabolic cart RER showed significant increases in metabolic levels, as a result of
glucose intake in healthy individuals in resting conditions (Figure 3). These results can be expected as cells
using more carbohydrates as fuel, produce more CO2 relative to the O2 consumption compared to cells
metabolizing fat. The ratio between the CO2 production and the O2 consumption in this process is known
as respiratory quotient (RQ) or RER. RQ and RER vary depending on the energy source of the cell
(carbohydrate vs. fat), and the acronyms are commonly used interchangeably [1–3]. In resting conditions,
oxygen consumption is fairly stable [11,12], meaning that participants’ changes in the RQ are due to
changes in CO2 production. This is the concept underlying the Lumen device, enabling it to track changes
in metabolic fuel utilization. For that reason, it was important to ensure that participants in this study
were at rest before and during their measurements.
Reduced major axis regression revealed an agreement between RER and Lumen %CO2 (Figure 4). This
analysis enables to test for agreement between methods with different units and verified the validity of
the Lumen device with the gold standard metabolic cart. The Bland-Altman plot showed a mean bias of
3.505, and when used in conjunction with a significant linear model of difference scores compared to
mean values , support the agreement between the two methods despite measuring in different units
(Figure 5). These results demonstrate the ability of Lumen to provide comparable results to the metabolic
cart in assessing the metabolic fuel utilization.
Furthermore, the results from the simple linear regression predicting Lumen %CO2 using RER values,
suggest that while there is measurement agreement between the Lumen %CO2 and RER, the proportion
of explained variance remains low (Figure 6). Thus, Lumen can be seen to be an effective instrument for
monitoring relative, individual changes in metabolic responses (within-subject consistency), rather than a
substitute for the metabolic cart (between-subject precision).
Evidence suggests that the assessment of RER can be a benefit for multiple conditions, such as nourishing,
diabetes prevention, weight management, physical activity, and healthy lifestyle [14,15]. It has previously
been shown that RER could be a prognostic marker of weight loss and a predictor of weight gain [16,17].
Moreover, minute-by-minute RER corresponded directly to intensity of exercise, and slopes of RER were
different in response to different dietary treatments . However, although RER is currently the
preferred method for determining metabolic fuel, it is a costly, time consuming, uncomfortable and an
impractical tool for assessing metabolic activity over time and for real time day to day usage. In contrast,
the Lumen device is small, relatively cheap, mobile, user-specific, delivers the outcome immediately to
the user and enables taking real time decisions.
This study is the first to show agreement between Lumen %CO2 and RER. However, it is important to note
that participants in this study were young (mean age: 22.4 y) and healthy individuals. With increasing age,
metabolism changes as can be seen in metabolic cart studies [19–21]. Future studies will need to examine
whether RER metabolic cart levels correspond to Lumen CO2% levels in older subjects and those with
Another limitation is that unlike the metabolic cart, the Lumen device does not measure oxygen
consumption. Accordingly, the Lumen measurement should be performed under resting conditions with
subsequent stable VO2, allowing the correct interpretation of changes of %CO2 as changes in metabolic
In addition, results from this study show high peak of blood glucose levels 45 minutes after glucose intake
(5 minutes after the 3rd drink), whereas both RER and Lumen %CO2 showed a more moderate increase in
levels. Therefore, it is possible that the metabolic cart and Lumen measurements were performed too
early, as it may be that in some of our participants the peak glucose occurred later than at 45 minutes,
thus not yet fully metabolized .
In summary, Lumen® can provide valid information regarding an individual's resting metabolic state, and
in agreement with results from metabolic cart. Unlike the metabolic cart, Lumen measurement can be
performed anywhere, without the need for a specialized lab, equipment, and technical support. The
capability of taking metabolic measurements continuously can provide numerous insights about the
metabolic state of an individual, for further scientific knowledge and understanding about metabolism
Conflict of Interest
SY and MM are employees of Metaflow Ltd., and contributed to the design and analysis of the study as
well as the preparation of the manuscript. The other authors declare no conflicts of interest.
This study was approved by the University’s Institutional Review Board for Human Subjects, and written
informed consent was obtained from each participant before testing.
KAL analyzed the data and prepared the manuscript
SY analyzed the data and prepared the manuscript
RA coordinated the project and collected the data
JO coordinated the project and collected the data
JRB reviewed and edited the manuscript
MM conceived, designed, and supervised the study as well as reviewed and edited the manuscript
MK conceived, designed, and supervised the study as well as reviewed and edited the manuscript
All authors approved the manuscript before submission
This work was supported by Metaflow Ltd.
We would like to thank the participants for their time in taking part in this study, and the Lumen team for
their support. We would like to acknowledge Casey Curl for his work in setting up the protocols and
procedures during preliminary testing.
1. Brooks GA, Fahey TD, Baldwin KM. Exercise physiology: human bioenergetics and its applications.
Exerc Physiol Hum Bioenerg its Appl. Independently Published; 2019.
2. Benedict FG, Cathcart EP. Muscular Work, a Metabolic Study with Special Reference to the
Efficiency of the Human Body as a Machine. Carnegie Inst Washingt. Carnegie institution of
3. McClave SA, Lowen CC, Kleber MJ, McConnell JW, Jung LY, Goldsmith LJ. Clinical use of the
respiratory quotient obtained from indirect calorimetry. J Parenter Enter Nutr 2003;27(1):21–26.
4. Moon JK, Butte NF. Combined heart rate and activity improve estimates of oxygen consumption
and carbon dioxide production rates. J Appl Physiol 1996;81(4):1754–1761. [doi:
5. Tsoukias NM, Tannous Z, Wilson AF, George SC, Nikolaos M, Tannous Z, Wil- AF, George SC.
Single-exhalation profiles of NO and CO 2 in humans : effect of dynamically changing flow rate. J
Appl Physiol 1998;85(2):642–652.
6. Brace RA. Fitting straight lines to experimental data. Am J Physiol 1977;233(3).
7. Ludbrook J. Linear regression analysis for comparing two measurers or methods of
measurement: But which regression? Clin Exp Pharmacol Physiol 2010;37(7):692–699. [doi:
8. Ludbrook J. Comparing measurment methods. Clin Exp Pharmacol Physiol 1997;24(July
9. Giavarina D. Understanding Bland Altman Analysis. Biochem Medica 2015;25(2):141–51.
10. Bland MJ, Altman DG. Statistical Methods for Assessing Agreement Between Two Methods of
Clinical Measurement. Lancet [Internet] 1986;327(8476):307–310. PMID:2868172
11. Spurr GB, Prentice AM, Murgatroyd PR, Goldberg GR, Reina JC, Christman NT. Energy expenditure
from minute-by-minute heart-rate recording: Comparison with indirect calorimetry. Am J Clin
Nutr 1988;48(3):552–559. PMID:3414570
12. Green JA. The heart rate method for estimating metabolic rate: Review and recommendations.
Comp Biochem Physiol - A Mol Integr Physiol Elsevier Inc.; 2011 Mar 1;158(3):287–304. [doi:
13. Woo P, Murthy G, Wong C, Hursh B, Chanoine JP, Elango R. Assessing resting energy expenditure
in overweight and obese adolescents in a clinical setting: Validity of a handheld indirect
calorimeter. Pediatr Res 2017;81(1):51–56. [doi: 10.1038/pr.2016.182]
14. Ferrannini E. The theoretical bases of indirect calorimetry: A review. Metabolism. Elsevier; 1988.
p. 287–301. [doi: 10.1016/0026-0495(88)90110-2]
15. Ramos-Jiménez A, Hernández-Torres RP, Torres-Durán P V., Romero-Gonzalez J, Mascher D,
Posadas-Romero C, Juárez-Oropeza MA. The Respiratory Exchange Ratio is Associated with
Fitness Indicators Both in Trained and Untrained Men: A Possible Application for People with
Reduced Exercise Tolerance. Clin Med Circ Respirat Pulm Med 2008;2:CCRPM.S449.
16. Valtueña S, Salas-Salvadó J, Lorda PG. The respiratory quotient as a prognostic factor in weight-
loss rebound. Int J Obes 1997;21(9):811–817. [doi: 10.1038/sj.ijo.0800480]
17. Zurlo F, Lillioja S, Puente AED, Nyomba BL, Raz I, Saad MF, Swinburn BA, Knowler WC, Bogardus
C, Ravussin E. Low ratio of fat to carbohydrate oxidation as predictor of weight gain: Study of 24-
h RQ. Am J Physiol - Endocrinol Metab 1990;259(5 22-5). PMID:2240203
18. Gribok A, Leger JL, Stevens M, Hoyt R, Buller M, Rumpler W. Measuring the short-term substrate
utilization response to high-carbohydrate and high-fat meals in the whole-body indirect
calorimeter. Physiol Rep 2016;4(12). [doi: 10.14814/phy2.12835]
19. Riera CE, Dillin A. Tipping the metabolic scales towards increased longevity in mammals. Nat Cell
Biol [Internet] Nature Publishing Group; 2015 Mar 27 [cited 2020 May 4];17(3):196–203. [doi:
20. Short KR, Vittone JL, Bigelow ML, Proctor DN, Nair KS. Age and aerobic exercise training effects
on whole body and muscle protein metabolism. Am J Physiol - Endocrinol Metab 2004;286(1 49-
1):92–101. [doi: 10.1152/ajpendo.00366.2003]
21. Gupta RD, Ramachandran R, Venkatesan P, Anoop S, Joseph M TN. Indirect Calorimetry: From
Bench to Bedside. Indian J Endocrinol Metab | 2017;21(4):594–599. [doi:
22. Eyth E, Basit H, Smith CJ. Glucose Tolerance Test. StatPearls Publishing; 2019. PMID:30422510