This paper is in the following e-collection/theme issue:

mHealth for Wellness, Behavior Change and Prevention (3314) Wearable Devices and Sensors (1133) mHealth for Data Collection and Research (1196) Digital Biomarkers and Digital Phenotyping (284) mHealth for Symptom and Disease Monitoring, Chronic Disease Management (1732)

Published on in Vol 13 (2025)

Preprints (earlier versions) of this paper are available at https://preprints.jmir.org/preprint/64604, first published .
The Association of Previous Day Carbohydrate Consumption With Fasted, Exhaled Carbon Dioxide in Lumen Users: Retrospective Real-World Study

The Association of Previous Day Carbohydrate Consumption With Fasted, Exhaled Carbon Dioxide in Lumen Users: Retrospective Real-World Study

The Association of Previous Day Carbohydrate Consumption With Fasted, Exhaled Carbon Dioxide in Lumen Users: Retrospective Real-World Study

Authors of this article:

Shlomo Yeshurun1 Author Orcid Image ;   Tomer Cramer1 Author Orcid Image ;   Daniel Souroujon1, 2 Author Orcid Image ;   Merav Mor1 Author Orcid Image
Article Authors Cited by Tweetations Metrics

    Original Paper

    1Metaflow Ltd, Tel Aviv, Israel

    2School of Public Health, Tel Aviv University, Tel Aviv, Israel

    Corresponding Author:

    Shlomo Yeshurun, PhD

    Metaflow Ltd

    HaArba'a Street 30

    Tel Aviv

    Israel

    Phone: 972 37684062

    Email: shlomi@lumen.me


    Background: The increasing prevalence of obesity and related metabolic disorders has highlighted the need for innovative tools to monitor and manage metabolic health. The Lumen device offers a novel approach to assess the metabolic state through exhaled carbon dioxide (CO2) measurements, providing real-time feedback via a mobile app. This app-driven experience allows users to track their metabolic state and receive personalized nutrition and lifestyle recommendations, potentially supporting long-term metabolic health improvements.

    Objective: This study aimed to investigate the association between the previous day’s carbohydrate consumption with fasted, exhaled CO2 levels in female and male Lumen users while also examining the influence of fasting duration, BMI, and age.

    Methods: We conducted a retrospective, observational study using deidentified data from 48,058 Lumen users, comprising 707,372 fasted sessions. Separate linear mixed models were fitted for female (n=520,269 sessions) and male (n=187,103 sessions) users due to observed sex differences in metabolism. User ID was included as a random effect to account for repeated measures. The models analyzed the relationship between fasted %CO2 levels and reported carbohydrate intake, fasting duration, BMI, and age.

    Results: Higher reported carbohydrate intake from the previous day was significantly associated with increased morning %CO2 levels in both male and female users (β=.032; P<.001; Cohen d=0.0691 in women and β=.024; P<.001; Cohen d=0.0534 in men), while a longer fasting duration was linked to decreased %CO2 levels in both sexes (β=−.017; P<.001 for both and Cohen d=−0.0374 in women and Cohen d=−0.0382 in men). A higher BMI was associated with elevated %CO2 levels in both sexes (β=.018; P<.001; Cohen d=0.0390 in women and β=.017; P<.001; Cohen d=0.0384 in men). Age had a statistically significant but modest effect in women (β=.008; P<.001; Cohen d=0.0180), whereas the effect size was minimal and did not meet the stricter significance threshold in men (β=.001; P=.02; Cohen d=0.0016). Cohen d values indicated that reported carbohydrate intake had the strongest effect size, while fasting duration and BMI had relatively smaller effects in both models.

    Conclusions: The Lumen device is able to detect changes in fasted %CO2 levels based on the previous day’s reported carbohydrate intake and fasting duration with sex-specific metabolic responses. These findings highlight the potential of Lumen as a personalized metabolic health monitoring tool, providing insights into the influence of dietary intake and fasting on metabolic state. Future research should investigate the hormonal and physiological mechanisms contributing to the observed sex differences and assess the long-term impact of app-guided metabolic feedback on user behaviors and metabolic health outcomes.

    JMIR Mhealth Uhealth 2025;13:e64604

    doi:10.2196/64604

    Keywords

    Lumen; fasting; carbohydrate; mHealth; BMI; metabolism 


    Background

    There has been a dramatic increase in the prevalence of obesity in recent years, posing significant health challenges globally [1,2]. Obesity is closely linked to an elevated risk of chronic diseases, including type 2 diabetes and cardiovascular diseases, which are leading causes of mortality worldwide [3,4]. Despite efforts to promote lifestyle changes and reduce caloric intake, sustained weight management remains challenging for many individuals, highlighting the need for innovative and effective interventions [5,6].

    It has been proposed that modifying the macronutrient composition of the diet could induce weight loss [7,8]. For instance, diets low in carbohydrates or high in protein have shown promise in promoting satiety and supporting weight loss [9-11]. By modulating macronutrient intake, individuals may experience improved metabolic flexibility, which is the body’s ability to efficiently switch between burning carbohydrates and fats for energy, enabling better regulation of blood glucose levels and lipid metabolism [12,13]. Metabolic flexibility can be assessed by measuring the respiratory exchange ratio (RER) from metabolic cart measurements, where a lower RER during fasted states indicates greater reliance on fat oxidation over carbohydrate oxidation [13]. Short-term fasting has been shown to decrease RER values, reflecting the shift toward lipid metabolism and improved metabolic flexibility [14,15]. Metabolic flexibility is also linked to obesity, as individuals with a higher BMI are less flexible and more likely to experience metabolic disorders [16-19]. Therefore, enhancing metabolic flexibility through dietary adjustments may offer a viable strategy for mitigating the adverse effects of obesity.

    Time-restricted eating is another recently popular diet, in which meals are limited to a specific period during the day. Several studies have shown that it is beneficial in weight loss, as well as other health benefits such as improved insulin sensitivity, reduced inflammation, and higher metabolic flexibility [20-22]. In addition, it may help to lower the risk of chronic diseases such as heart disease and diabetes [23]. Time-restricted eating has also been shown to decrease RER values during the fasted state, indicating a shift toward greater fat oxidation [24]. This reduction in RER during fasting periods suggests an enhanced capacity to use fat as a fuel source, which may contribute to the beneficial metabolic effects of intermittent fasting regimens [25].

    Mobile health technologies, such as the Lumen breath analyzer, present innovative opportunities for real-time monitoring of metabolic responses to dietary changes [26,27]. Moreover, we recently showed that increased carbohydrate consumption in meals was associated with increased postprandial %CO2 response as measured by Lumen [28]. Similarly, applying continuous glucose monitoring (CGM) system with a mobile app was associated with improved glycemic control [29]. Such technologies empower users to make informed decisions about their diet and lifestyle, thereby potentially enhancing long-term health outcomes [30-33].

    Objective

    This study aimed to investigate whether fasted %CO2 measurements are associated with the previous day’s carbohydrate intake, fasting duration as logged by Lumen users, and users’ age, sex, and BMI.


    Study Design

    This is a retrospective observational study based on deidentified data collected from users of the Lumen device and app.

    Participants

    Participants’ data were collected retrospectively between September 16, 2023, and March 14, 2024. All users in the analysis were aged ≥18 years. As only 150 users in the database were found to be underweight (BMI ≤18.5), they were removed from the analysis.

    Data Sources

    During the onboarding process, users had to specify their sex (assigned at birth), date of birth, height, and weight. The Lumen app allows users to perform a variety of optional measurements with the Lumen device (MetaFlow Ltd), such as morning fasted, as well as pre- and postmeal measurements and a bedtime measurement before sleeping. The app guides the user through each phase of the breathing maneuver, and exhaled CO2 levels are obtained, as previously described [26]. The app provides nutritional recommendations based on the user’s personal preferences and their morning %CO2 measurements regarding how much macronutrients should be consumed during the day. In addition, with each morning measurement, the app asks users about the number of portions of carbohydrates they consumed the day before (1 portion=15 g) and when was the last time they consumed any food. An example of the morning questions in the app is shown in Figure 1. For this study, only fasted measurements where users reported both their carbohydrate intake and the time they stopped eating the previous day before were included. To avoid discrepancies, we removed the top and bottom 2.5% of data points for both the amount of carbohydrate portions consumed the previous day, and the duration of fasting hours, effectively eliminating outliers from both extremes. Moreover, in instances where more than half of a user’s measurements were removed for these reasons, that user’s entire dataset was consequently excluded from the analysis.

    ‎
    Figure 1. Lumen mobile app screenshots (A and B) detailing the user input asked for following a fasted %CO2 measurement with the Lumen device. (C) An illustration of the Lumen device.

    Statistical Analysis

    Data were analyzed using R Studio (version 2023.12.1; Posit PBC), and all statistical analyses were conducted with the R (version 4.3.3; R Core Team) programming language, using custom scripts and the lme4 package [34].

    Linear mixed models were fitted separately for women and men to explore how the previous day’s carbohydrate consumption, fasting duration, and personal characteristics (age and BMI) influence fasted %CO2 levels. Separating the sexes for analysis was based on well-documented metabolic differences and our previous work, which identified distinct %CO2 patterns in female Lumen users related to menstrual cycles and menopause [35]. Users’ unique IDs were included as a random effect to account for individual variability. Model fit was evaluated using Akaike Information Criterion, Bayesian Information Criterion, and R² values, with detailed results provided in Table S1 in Multimedia Appendix 1. Full random effects results and diagnostic checks are also presented in Table S1 and Figure S1 in Multimedia Appendix 1. Effect sizes were quantified using Cohen d to assess the relative strength of each predictor variable in explaining %CO2 variation.

    Among all performed analysis, a 2-sided P≤.01 was considered statistically significant, following Maier and Lakens [36] and Wulff and Taylor [37] to mitigate inflated type I errors in large datasets and ensure robust findings.

    Ethical Considerations

    This study was determined to be exempt from institutional review board (IRB) under category 2, as detailed in 45 CFR 46.104(d) and the standard operating procedure of the Biomedical Research Alliance of New York (BRANY), by the BRANY Social, Behavioral, and Educational Research IRB on May 9, 2023 (BRANY IRB File 23-119-1476). Exemption was granted because the study involved the secondary use of deidentified data that were originally collected for purposes other than this research, ensuring that participants could not be identified. All user data were fully anonymized to minimize the risk of violating participants’ privacy, and no identifiable information was accessible to the researchers. As such, informed consent was not required, and no compensation was given to the participants.


    Participants

    Overall, 48,058 users have completed 707,372 fasted sessions reporting their carbohydrate intake from the previous day. These sessions were included in the final analysis, with a median of 8 (IQR 6-10) sessions per user.

    Descriptive Characteristics

    Among the users, 71.24% (34,238/48,058) were women. Female participants were slightly older than their male counterparts, with a mean age of 49.9 (SD 10) years compared to 48.1 (SD 10.2) years, and had a slightly lower BMI, averaging 29.2 (SD 6) kg/m2 compared to 30.4 (SD 5.3) kg/m2. A summary of the participants’ characteristics is presented in Table 1.

    Across 707,372 morning measurements, users reported consuming an average of 7.9 (SD 3) portions of carbohydrates the previous day. Women consumed an average of 7.5 (SD 2.7) portions, while men consumed 9.1 (SD 3.5) portions. The average fasting duration was 11.3 (SD 1.8) hours, with women fasting for 11.3 (SD 1.7) hours and men for 11.1 (SD 1.8) hours.

    Table 1. Sample characteristics of users (n=48,058).
    CharacteristicsFemaleMaleOverall
    Sample, n (%)34,238 (71.24)13,820 (28.76)48,058 (100)
    Age (y), mean (SD)49.9 (10)48.1 (10.2)50.0 (9.7)
    BMI (kg/m2), mean (SD)29.2 (6)30.4 (5.3)28.9 (5.6)

    Evaluation Outcomes

    Figure 2 presents the relationship between fasted %CO2 levels and previous day’s reported carbohydrates (A and B) and fasting duration (C and D), categorized by BMI (A and C) and sex (B and D). As expected, increased reported carbohydrate consumption in the previous day resulted in increased %CO2, and longer fasting duration resulted in decreased %CO2. In addition, users with obesity showed higher %CO2 than users with overweight, which in turn had higher %CO2 than those with healthy BMI. Moreover, female users displayed higher %CO2 than male users, with a mean morning %CO₂ of 4.60% (SD 0.47) in women and 4.56% (SD 0.46%) in men.

    ‎
    Figure 2. Fasted %CO2 measurements in response to different parameters (mean and 95% CI): (A) previous day’s reported carbohydrate consumption, categorized by BMI groups (healthy, overweight, and obese); (B) previous day’s carbohydrate consumption, categorized by sex (male and female); (C) fasting duration, categorized by BMI groups; and (D) fasting duration, categorized by sex.

    The relationship between fasted %CO2 levels and previous day’s carbohydrates, fasting duration, and users’ personal information as recorded in the Lumen app was examined using separate linear mixed models for women (Table 2) and men (Table 3). User ID was modeled as a random effect to account for interindividual variability (see full random effects results in Table S1 in Multimedia Appendix 1). In both sexes, higher BMI and reported carbohydrate intake were associated with increased %CO2 levels (reported carbohydrate intake: β=.032; P<.001; Cohen d=0.0691 in women and β=.024; P<.001; Cohen d=0.0534 in men and BMI: β=.018; P<.001; Cohen d=0.0390 in women and β=.017; P<.001; Cohen d=0.0384 in men), while prolonged fasting was linked to decreased %CO2 levels (β=−.017; P<.001 for both and Cohen d=−0.0374 in women and Cohen d=−0.0382 in men), corroborating the observations in Figure 2. Cohen d values suggest that carbohydrate intake had the strongest effect size, while fasting duration and BMI had relatively smaller effects in both models. Age had a small but statistically significant effect only in women (β=.008; P<.001; Cohen d=0.0180).

    Both models demonstrated good fit, with diagnostic checks supporting their validity. Further details, including Akaike Information Criterion, Bayesian Information Criterion, R² values, and full diagnostic results, are provided in Table S1 and Figure S1 in Multimedia Appendix 1.

    Table 2. Determinants of %CO2 in a fasted state in women (n=520,269).
    Variablesβ (95% CI)z statisticsCohen dP value
    Intercept3.671 (3.642 to 3.700)253.688.0015<.001
    Previous day’s carbohydrates (portions)0.032 (0.031 to 0.032)133.000.0691<.001
    Fasting duration (h)−0.017 (−0.0018 to −0.017)−55.28−0.0374<.001
    BMI (kg/m2)0.018 (0.017 to 0.019)53.110.0390<.001
    Age (y)0.008 (0.008 to 0.009)40.540.0180<.001
    Table 3. Determinants of %CO2 in a fasted state in men (N=187,103).
    Variablesβ (95% CI)z statisticsCohen dP value
    Intercept4.010 (3.963 to 4.057)165.5818.8970<.001
    Previous day’s carbohydrates (portions)0.024 (0.023 to 0.025)85.7410.0534<.001
    Fasting duration (h)−0.017 (−0.0018 to −0.016)−35.173−0.0382<.001
    BMI (kg/m2)0.017 (0.016 to 0.019)29.2570.0384<.001
    Age (y)0.001 (0.000 to 0.002)2.3170.0016.021

    Principal Findings

    This study provides evidence that the Lumen device can detect changes in morning %CO2 levels in relation to reported carbohydrate intake from the previous day, with similar metabolic responses observed in male and female Lumen users, apart from a notable age effect in women (Tables 2 and 3). The response is also influenced by fasting duration and BMI to a lesser extent (Figure 2).

    Higher reported carbohydrate intake was associated with increased morning %CO2 in both women and men (Figure 2) aligning with previous studies showing similar results for RER values measured by metabolic carts [13,16]. In addition, prolonged fasting duration was linked to decreased %CO2, highlighting the influence of fasting on metabolic processes, which is consistent with findings on time-restricted feeding [24].

    The observed shifts in morning %CO₂ values occurred within a physiologically plausible range; the mean morning %CO₂ was 4.60% (SD 0.47%) in women and 4.56% (SD 0.46%) in men. Although the effect sizes were small, even modest changes on the order of 0.03% to 0.05% may reflect meaningful shifts in metabolic substrate use, particularly in large-scale, real-world datasets, in accordance with our previous retrospective analysis of female users [35]. In addition, in a controlled study, fasted %CO₂ levels increased by approximately 0.11% when comparing high- versus low-carbohydrate diets over a 7-day period, demonstrating that even modest shifts can reflect meaningful metabolic adaptations [27]. These results support the physiological relevance of such variations in exhaled CO₂, particularly in relation to carbohydrate intake and metabolic flexibility.

    As shown in our previous research, higher BMI was associated with increased %CO2 levels [28], suggesting ameliorated metabolic flexibility in individuals with obesity. While most associations were consistent across both sexes, age demonstrated a stronger effect in female Lumen users compared to men, as older age was correlated with elevated %CO2 levels in women. This may be due to differences in menstrual cycle phases and menopause status, as previously observed in Lumen users [35], where eumenorrheic women exhibited lower %CO2 levels compared to postmenopausal women, suggesting hormonal influence on exhaled %CO2.

    Comparison With Prior Work

    The findings of our study align with and expand upon previous research in the field of metabolic health and nutrition. Our observation that higher reported carbohydrate intake is associated with increased morning %CO2 is consistent with findings from studies using traditional RER measurements. For instance, Galgani et al [13] demonstrated that carbohydrate overfeeding led to increased RER values, indicating a shift toward carbohydrate oxidation. In addition, this retrospective analysis showed similar results to a previous prospective study using the Lumen device. Roberts et al [27] observed significant changes in exhaled %CO2 in response to short-term low and high carbohydrate diets, where high carbohydrate diet resulted in higher fasted %CO2 levels. Our study adds to this by demonstrating that these effects are observable in a larger, real-world setting, with variations in carbohydrate intake from the previous day significantly influencing morning %CO2 measurements.

    The inverse relationship we observed between fasting duration and morning %CO2 levels aligns with established literature on the metabolic effects of fasting. Ravussin et al [24] demonstrated that time-restricted feeding increased fat oxidation and decreased RER values in addition to improved metabolic flexibility in overweight adults. Similarly, Vinales et al [15] showed that fasting led to decreased RER values from the metabolic chamber, while high carbohydrate intake resulted in increased RER. As longer overnight fasting was previously associated with reduced BMI [38], this study suggests that Lumen could be a valuable tool for monitoring the metabolic effects of various fasting regimens which might assist in weight loss.

    Our study identified significant associations between %CO2 levels and various demographic factors, including BMI, age, and sex. The positive correlation between BMI and %CO2 levels is consistent with previous research on metabolic flexibility. For instance, Begaye et al [14] found that individuals with higher BMI exhibited impaired metabolic flexibility, which could explain the higher %CO2 levels observed in our study. In addition, our previous retrospective analysis demonstrated the relationship between low metabolic flexibility in users with obesity [28].

    The observed associations between carbohydrate intake, fasting duration, and exhaled CO₂ may be explained by the role of insulin in substrate oxidation, as carbohydrate intake stimulates insulin secretion, promoting carbohydrate use while suppressing fat oxidation [13,19]. Similarly, the reduction in %CO₂ with prolonged fasting reflects a metabolic shift from carbohydrate to fat oxidation, indicative of improved metabolic flexibility [14,16]. Hormonal responses to fasting, including increased glucagon and cortisol and reduced insulin, further modulate substrate use [24,39]. In women, progesterone fluctuations during the menstrual cycle may also contribute to metabolic flexibility, as higher progesterone levels have been linked to increased lipid oxidation and lower RER values, which could partly explain the observed sex differences in age-related %CO₂ patterns [35,40,41]. Future research should further investigate these hormonal influences to better understand individual metabolic responses.

    The observed sex-based differences in our study are particularly interesting in light of our recent work by Cramer et al [35], which explored the relationship between the menstrual cycle and menopause to exhaled CO2 in Lumen users. This study showed cyclical %CO2 patterns in eumenorrheic women, with lower values compared to menopausal women, suggesting a hormonal influence on metabolic fuel use. Consistently, our findings indicate that age was significantly associated with %CO₂ in women but had a negligible effect in men. Given that most Lumen users analyzed in this study are women, this could explain the stronger association of age with fasted %CO₂ levels in the female cohort. Future research should aim to elucidate the underlying hormonal and physiological mechanisms driving these sex-specific responses and their potential implications for personalized nutrition and health strategies.

    Longitudinal investigations into the relationship between exhaled CO2 levels and weight loss outcomes could provide crucial insights into the potential of Lumen device for weight management. These studies could explore whether fluctuations in morning %CO2 levels serve as correlates of long-term weight loss success. In addition, the applicability of the device should be assessed in diverse populations, including individuals with metabolic disorders such as type 2 diabetes or polycystic ovary syndrome, as well as athletes and older adults, to determine whether metabolic responses differ across various health and lifestyle conditions. Future research should also investigate the integration of Lumen with other health monitoring technologies, such as CGMs or wearable activity trackers, to provide a more comprehensive view of metabolic health. Furthermore, future research should delve into the impact of additional parameters on exhaled CO2 levels. For instance, examining both the acute and chronic effects of physical activity on %CO2 measurements could yield valuable data as it was previously shown to influence RER levels [42]. Given the established links between sleep and metabolic health [43], investigating how sleep duration and quality influence %CO2 levels would also be beneficial. Similarly, exploring the role of stress levels is important, as cortisol is known to affect metabolic processes and was found to be elevated during prolonged fasting [39]. Ultimately, this research could pave the way for more personalized and effective strategies in metabolic health management.

    Our study contributes significantly to the growing field of mobile health (mHealth) and ubiquitous health technologies. The use of the Lumen device as a portable metabolic measurement tool aligns with the broader trend of leveraging consumer-grade devices for health monitoring and management. Similar to how CGMs have revolutionized diabetes management [29], our findings suggest that breath-based metabolic monitoring could play a crucial role in personalized nutrition and weight management strategies. The large-scale, real-world nature of our dataset is particularly valuable in the context of mHealth research, as it provides insights into how these technologies perform in everyday settings, outside of controlled laboratory environments. This approach is consistent with recent mHealth studies that have demonstrated the effectiveness of mobile apps in promoting health and managing diseases [30,31]. Furthermore, our study exemplifies the potential of mHealth devices to generate large datasets that can be used to identify patterns and personalize interventions, a key advantage noted in reviews of digital health interventions for weight loss [32]. By demonstrating the ability of Lumen device to detect metabolic changes in response to dietary habits and individual characteristics, our research contributes to the broader goal of empowering users to take greater responsibility for their health through accessible, user-friendly technologies [33].

    Limitations

    Despite the large dataset of real-world evidence that was used in this analysis, several limitations need to be mentioned. First, as this is a retrospective and observational study, causal relationships between variables cannot be definitively established. While the results provide valuable insights into associations between carbohydrate intake, fasting duration, and metabolic responses, future research using prospective or experimental designs is needed to strengthen causal inferences. Second, as this analysis is based on real-world data, the reported carbohydrate and the reported fasting duration may be miscalculated or incomplete (recall bias). Future studies should consider integrating objective dietary tracking technologies, such as food-tracking tools, to cross-validate self-reported data and reduce potential recall bias. Moreover, as only reported carbohydrate intake was considered in this analysis, it is difficult to determine if other factors in the meal might influence the %CO2 as well, such as fats and protein as well as the type of carbohydrate and the fiber composition. In addition, the macronutrient intake of Lumen users might not be representative of most typical diets, as the Lumen app guides them toward a specific diet, primarily low in carbohydrate, and with most of them on a weight loss journey. As this study has specific user characteristics, caution should be exercised in interpreting these conclusions for a broader audience. Future research should aim to validate findings in more diverse populations, including those with different dietary habits and health goals. Our model showed adequate but imperfect homoscedasticity (Figure S1 in Multimedia Appendix 1). Future studies could use generalized additive or robust regression techniques to address this limitation. Finally, fasted %CO2 levels might be affected by other confounding variables, including preexisting conditions, medication use, dietary restrictions, menstrual cycle, and lifestyle factors such as stress, sleep, and physical activity, which were not controlled for in this study. Future studies should control these variables through comprehensive data collection and statistical adjustments, or by designing prospective studies that account for these potential confounders.

    Conclusions

    In summary, this study demonstrates that the Lumen device can detect the impact of the previous day’s reported carbohydrate consumption on fasted %CO2 measurements in both female and male users. In addition, the device’s sensitivity to variations in fasting duration and BMI suggests its utility in providing personalized dietary recommendations. By accounting for individual differences in metabolic responses, Lumen can offer valuable insights into metabolic health and assist users in optimizing their carbohydrate intake and fasting routines.

    Acknowledgments

    This work was funded by MetaFlow Ltd. The authors would like to express their thanks to all app users who contributed their data to this study and the Lumen team for their support.

    Authors' Contributions

    SY, TC, and MM led the design and conceptualization of the study. MM was the principal supervisor of the study. SY led the curation of data and conducted data interpretation and formal analysis. SY wrote the first draft of the manuscript. TC and DS contributed to the data analysis. All authors critically revised the manuscript and gave their final approval.

    Conflicts of Interest

    MM is the chief scientific officer and cofounder of MetaFlow Ltd. DS is a consultant at MetaFlow. SY and TC are paid employees at MetaFlow. To promote transparency and mitigate potential conflicts of interest, the study was reviewed and approved by an independent institutional review board. Data analysis followed established methodologies, and statistical interpretations were based on predefined analytical approaches in accordance with existing literature. While the study was funded by MetaFlow Ltd, the decision to publish and the interpretation of findings were guided by scientific integrity, independent of the funder’s influence. All other authors declare no other conflicts of interest.

    Multimedia Appendix 1

    Supplementary materials including model fit statistics for the linear mixed models (Table S1) and residual diagnostic plots assessing homoscedasticity, normality, and distribution of residuals (Figure S1).

    DOC File , 32 KB
    1. NCD Risk Factor Collaboration (NCD-RisC). Trends in adult body-mass index in 200 countries from 1975 to 2014: a pooled analysis of 1698 population-based measurement studies with 19·2 million participants. Lancet. Apr 02, 2016;387(10026):1377-1396. [FREE Full text] [CrossRef] [Medline]
    2. Powell-Wiley TM, Poirier P, Burke LE, Després JP, Gordon-Larsen P, Lavie CJ, et al. American Heart Association Council on Lifestyle and Cardiometabolic Health; Council on CardiovascularStroke Nursing; Council on Clinical Cardiology; Council on EpidemiologyPrevention;Stroke Council. Obesity and cardiovascular disease: a scientific statement from the American Heart Association. Circulation. May 25, 2021;143(21):e984-1010. [FREE Full text] [CrossRef] [Medline]
    3. Joseph JJ, Deedwania P, Acharya T, Aguilar D, Bhatt DL, Chyun DA, et al. American Heart Association Diabetes Committee of the Council on LifestyleCardiometabolic Health; Council on Arteriosclerosis‚ ThrombosisVascular Biology; Council on Clinical Cardiology;Council on Hypertension. Comprehensive management of cardiovascular risk factors for adults with type 2 diabetes: a scientific statement from the American Heart Association. Circulation. Mar 2022;145(9):e722-e759. [FREE Full text] [CrossRef] [Medline]
    4. Censin JC, Peters SA, Bovijn J, Ferreira T, Pulit SL, Mägi R, et al. Causal relationships between obesity and the leading causes of death in women and men. PLoS Genet. Oct 2019;15(10):e1008405. [FREE Full text] [CrossRef] [Medline]
    5. Wadden TA, Tronieri JS, Butryn ML. Lifestyle modification approaches for the treatment of obesity in adults. Am Psychol. 2020;75(2):235-251. [FREE Full text] [CrossRef] [Medline]
    6. MacLean PS, Wing RR, Davidson T, Epstein L, Goodpaster B, Hall KD, et al. NIH working group report: innovative research to improve maintenance of weight loss. Obesity (Silver Spring). Jan 2015;23(1):7-15. [FREE Full text] [CrossRef] [Medline]
    7. Ge L, Sadeghirad B, Ball GD, da Costa BR, Hitchcock CL, Svendrovski A, et al. Comparison of dietary macronutrient patterns of 14 popular named dietary programmes for weight and cardiovascular risk factor reduction in adults: systematic review and network meta-analysis of randomised trials. BMJ. Apr 01, 2020;369:m696. [FREE Full text] [CrossRef] [Medline]
    8. Austin GL, Ogden LG, Hill JO. Trends in carbohydrate, fat, and protein intakes and association with energy intake in normal-weight, overweight, and obese individuals: 1971-2006. Am J Clin Nutr. Apr 2011;93(4):836-843. [FREE Full text] [CrossRef] [Medline]
    9. Halford JC, Harrold JA. Satiety-enhancing products for appetite control: science and regulation of functional foods for weight management. Proc Nutr Soc. May 2012;71(2):350-362. [FREE Full text] [CrossRef] [Medline]
    10. Kim JY. Optimal diet strategies for weight loss and weight loss maintenance. J Obes Metab Syndr. Mar 30, 2021;30(1):20-31. [FREE Full text] [CrossRef] [Medline]
    11. Wycherley TP, Moran LJ, Clifton PM, Noakes M, Brinkworth GD. Effects of energy-restricted high-protein, low-fat compared with standard-protein, low-fat diets: a meta-analysis of randomized controlled trials. Am J Clin Nutr. Dec 2012;96(6):1281-1298. [FREE Full text] [CrossRef] [Medline]
    12. Russell WR, Baka A, Björck I, Delzenne N, Gao D, Griffiths HR, et al. Impact of diet composition on blood glucose regulation. Crit Rev Food Sci Nutr. 2016;56(4):541-590. [FREE Full text] [CrossRef] [Medline]
    13. Galgani JE, Moro C, Ravussin E. Metabolic flexibility and insulin resistance. Am J Physiol Endocrinol Metab. Nov 2008;295(5):E1009-E1017. [FREE Full text] [CrossRef] [Medline]
    14. Begaye B, Vinales KL, Hollstein T, Ando T, Walter M, Bogardus C, et al. Impaired metabolic flexibility to high-fat overfeeding predicts future weight gain in healthy adults. Diabetes. Feb 2020;69(2):181-192. [FREE Full text] [CrossRef] [Medline]
    15. Vinales KL, Schlögl M, Piaggi P, Hohenadel M, Graham A, Bonfiglio S, et al. The consistency in macronutrient oxidation and the role for epinephrine in the response to fasting and overfeeding. J Clin Endocrinol Metab. Jan 01, 2017;102(1):279-289. [FREE Full text] [CrossRef] [Medline]
    16. Goodpaster BH, Sparks LM. Metabolic flexibility in health and disease. Cell Metab. May 02, 2017;25(5):1027-1036. [FREE Full text] [CrossRef] [Medline]
    17. Galgani JE, Fernández-Verdejo R. Pathophysiological role of metabolic flexibility on metabolic health. Obes Rev. Feb 2021;22(2):e13131. [CrossRef] [Medline]
    18. Palmer BF, Clegg DJ. Metabolic flexibility and its impact on health outcomes. Mayo Clin Proc. Apr 2022;97(4):761-776. [FREE Full text] [CrossRef] [Medline]
    19. Smith RL, Soeters MR, Wüst RC, Houtkooper RH. Metabolic flexibility as an adaptation to energy resources and requirements in health and disease. Endocr Rev. Aug 01, 2018;39(4):489-517. [FREE Full text] [CrossRef] [Medline]
    20. Harvie M, Howell A. Potential benefits and harms of intermittent energy restriction and intermittent fasting amongst obese, overweight and normal weight subjects-a narrative review of human and animal evidence. Behav Sci (Basel). Jan 19, 2017;7(1):4. [FREE Full text] [CrossRef] [Medline]
    21. Gabel K, Hoddy KK, Haggerty N, Song J, Kroeger CM, Trepanowski JF, et al. Effects of 8-hour time restricted feeding on body weight and metabolic disease risk factors in obese adults: a pilot study. Nutr Healthy Aging. Jun 15, 2018;4(4):345-353. [FREE Full text] [CrossRef] [Medline]
    22. Patterson RE, Sears DD. Metabolic effects of intermittent fasting. Annu Rev Nutr. Aug 21, 2017;37:371-393. [FREE Full text] [CrossRef] [Medline]
    23. Patikorn C, Roubal K, Veettil SK, Chandran V, Pham T, Lee YY, et al. Intermittent fasting and obesity-related health outcomes: an umbrella review of meta-analyses of randomized clinical trials. JAMA Netw Open. Dec 01, 2021;4(12):e2139558. [FREE Full text] [CrossRef] [Medline]
    24. Ravussin E, Beyl RA, Poggiogalle E, Hsia DS, Peterson CM. Early time-restricted feeding reduces appetite and increases fat oxidation but does not affect energy expenditure in humans. Obesity (Silver Spring). Aug 2019;27(8):1244-1254. [FREE Full text] [CrossRef] [Medline]
    25. Antoni R, Johnston KL, Collins AL, Robertson MD. Intermittent v. continuous energy restriction: differential effects on postprandial glucose and lipid metabolism following matched weight loss in overweight/obese participants. Br J Nutr. Mar 2018;119(5):507-516. [CrossRef] [Medline]
    26. Lorenz KA, Yeshurun S, Aziz R, Ortiz-Delatorre J, Bagley JR, Mor M, et al. A handheld metabolic device (Lumen) to measure fuel utilization in healthy young adults: device validation study. Interact J Med Res. May 17, 2021;10(2):e25371. [FREE Full text] [CrossRef] [Medline]
    27. Roberts J, Dugdale-Duwell D, Lillis J, Pinto JM, Willmott A, Yeshurun S, et al. The efficacy of a home-use metabolic device (Lumen) in response to a short-term low and high carbohydrate diet in healthy volunteers. J Int Soc Sports Nutr. Dec 2023;20(1):2185537. [FREE Full text] [CrossRef] [Medline]
    28. Yeshurun S, Cramer T, Souroujon D, Mor M. The association of macronutrient consumption and BMI to exhaled carbon dioxide in lumen users: retrospective real-world study. JMIR Mhealth Uhealth. Apr 01, 2024;12:e56083. [FREE Full text] [CrossRef] [Medline]
    29. Kumbara AB, Iyer AK, Green CR, Jepson LH, Leone K, Layne JE, et al. Impact of a combined continuous glucose monitoring-digital health solution on glucose metrics and self-management behavior for adults with type 2 diabetes: real-world, observational study. JMIR Diabetes. Sep 11, 2023;8:e47638. [FREE Full text] [CrossRef] [Medline]
    30. Kang HS, Exworthy M. Wearing the future-wearables to empower users to take greater responsibility for their health and care: scoping review. JMIR Mhealth Uhealth. Jul 13, 2022;10(7):e35684. [FREE Full text] [CrossRef] [Medline]
    31. Iribarren SJ, Akande TO, Kamp KJ, Barry D, Kader YG, Suelzer E. Effectiveness of mobile apps to promote health and manage disease: systematic review and meta-analysis of randomized controlled trials. JMIR Mhealth Uhealth. Jan 11, 2021;9(1):e21563. [FREE Full text] [CrossRef] [Medline]
    32. Patel ML, Wakayama LN, Bennett GG. Self-monitoring via digital health in weight loss interventions: a systematic review among adults with overweight or obesity. Obesity (Silver Spring). Mar 2021;29(3):478-499. [CrossRef] [Medline]
    33. Deniz-Garcia A, Fabelo H, Rodriguez-Almeida AJ, Zamora-Zamorano G, Castro-Fernandez M, Alberiche Ruano MD, et al. WARIFA Consortium. Quality, usability, and effectiveness of mHealth apps and the role of artificial intelligence: current scenario and challenges. J Med Internet Res. May 04, 2023;25:e44030. [FREE Full text] [CrossRef] [Medline]
    34. Bates D, Mächler M, Bolker B, Walker S. Fitting linear mixed-effects models using lme4. J Stat Soft. 2015;67(1):33. [CrossRef]
    35. Cramer T, Yeshurun S, Mor M. Changes in exhaled carbon dioxide during the menstrual cycle and menopause. Digit Biomark. 2024;8(1):102-110. [FREE Full text] [CrossRef] [Medline]
    36. Maier M, Lakens D. Justify your alpha: a primer on two practical approaches. Adv Methods Pract Psychol Sci. May 30, 2022;5(2):251524592210803. [CrossRef]
    37. Wulff JN, Taylor L. How and why alpha should depend on sample size: a Bayesian-frequentist compromise for significance testing. Strateg Organ. Jan 06, 2024;22(3):550-581. [CrossRef]
    38. Xiao Q, Bauer C, Layne T, Playdon M. The association between overnight fasting and body mass index in older adults: the interaction between duration and timing. Int J Obes (Lond). Mar 2021;45(3):555-564. [FREE Full text] [CrossRef] [Medline]
    39. Nakamura Y, Walker BR, Ikuta T. Systematic review and meta-analysis reveals acutely elevated plasma cortisol following fasting but not less severe calorie restriction. Stress. 2016;19(2):151-157. [FREE Full text] [CrossRef] [Medline]
    40. Burger HG, Dudley EC, Robertson DM, Dennerstein L. Hormonal changes in the menopause transition. Recent Prog Horm Res. 2002;57:257-275. [CrossRef] [Medline]
    41. Rogan MM, Black KE. Dietary energy intake across the menstrual cycle: a narrative review. Nutr Rev. Jun 09, 2023;81(7):869-886. [FREE Full text] [CrossRef] [Medline]
    42. San-Millán I, Brooks GA. Assessment of metabolic flexibility by means of measuring blood lactate, fat, and carbohydrate oxidation responses to exercise in professional endurance athletes and less-fit individuals. Sports Med. Feb 2018;48(2):467-479. [FREE Full text] [CrossRef] [Medline]
    43. Reutrakul S, Van Cauter E. Sleep influences on obesity, insulin resistance, and risk of type 2 diabetes. Metabolism. Jul 2018;84:56-66. [CrossRef] [Medline]

    ‎
    BRANY: Biomedical Research Alliance of New York
    CGM: continuous glucose monitoring
    IRB: institutional review board
    RER: respiratory exchange ratio

    Edited by L Buis; submitted 23.07.24; peer-reviewed by Y Miao, P Shamanna; comments to author 04.02.25; revised version received 01.04.25; accepted 05.08.25; published 02.09.25.

    Copyright

    ©Shlomo Yeshurun, Tomer Cramer, Daniel Souroujon, Merav Mor. Originally published in JMIR mHealth and uHealth (https://mhealth.jmir.org), 02.09.2025.

    This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR mHealth and uHealth, is properly cited. The complete bibliographic information, a link to the original publication on https://mhealth.jmir.org/, as well as this copyright and license information must be included.