Skip to main content
NCBI home page
Journal of the International Society of Sports Nutrition logoLink to Journal of the International Society of Sports Nutrition
. 2025 Sep 16;22(1):2561670. doi: 10.1080/15502783.2025.2561670

Comparative effects of continuous glucose monitoring-informed and traditional interval-based carbohydrate refueling protocols on endurance exercise responses: an exploratory study

Eric Tsz-Chun Poon a,✉,*, Kin-Wong Lau b,*, Jaclyn Hei Tsang c, Xiaoyuan Zhang d, Waris Wongpipit a,e,f
PMCID: PMC12466181  PMID: 40957601

ABSTRACT

Background

Recent advancements in Continuous Glucose Monitoring (CGM) technology provide a novel solution for monitoring dynamic glucose fluctuations during endurance exercise, enabling personalized carbohydrate refueling strategies. However, direct evidence demonstrating its superiority over traditional refueling guidelines has been lacking. This exploratory study compared a CGM-informed protocol with a traditional interval-based (TRAD) carbohydrate refueling protocol on endurance exercise responses.

Methods

Twelve healthy, non-diabetic males (age = 32.3 ± 6.3 years) participated in two trials involving 75 minutes of continuous cycling at 50% peak power output, with the two carbohydrate refueling protocols implemented in a randomized, crossover order. In the TRAD protocol, participants received standard carbohydrate solutions at regular 15-min intervals, while replenishment timing in the CGM-informed protocol was guided by real-time glucose trends, particularly during declines.

Results

No significant differences were observed in heart rate, blood lactate, or perceived exertion between conditions (all p > 0.05). However, CGM-informed protocol displayed a lower coefficient of variation of glucose levels (8.82 ± 4.68%) than the TRAD protocol (11.95 ± 3.41%, Cohen’s d  = 0.76, p = 0.013), indicating a steadier blood glucose trend. The mean 75-minute interstitial glucose total area under the curve was significantly lower for the CGM-informed protocol (423.4 ± 39.1 mmol·min·L−1), compared to the TRAD protocol (448.6 ± 38.2 mmol·min·L−1, Cohen’s d = 0.65, p = 0.008).

Conclusion

While the CGM-informed protocol did not demonstrate beneficial effects on physiological responses over the TRAD protocol, it may aid in stabilizing glucose levels during prolonged endurance exercise. Further research is warranted to explore the applications of personalized carbohydrate refueling strategy based on CGM technology on athletic performance across various exercise protocols.

KEYWORDS: Carbohydrate availability, exercise performance, endurance training, sports nutrition

1. Introduction

Endurance sports predominantly rely on the aerobic energy system of the human body, with glucose serving as the primary fuel source, thereby maintaining a sustained energy supply for high performance [1–3]. There are several established carbohydrate refueling guidelines from internationally recognized organizations including the American College of Sports Medicine (ACSM) and the International Society of Sports Nutrition (ISSN), which are commonly adopted by sports professionals and athletes [4,5]. The current ACSM guideline [5] recommends refueling 30 to 60 grams of carbohydrates per hour during endurance exercise lasting 1 to 2.5 hours. Similarly, the ISSN guideline [4] recommends refueling 30 to 60 grams of carbohydrates per hour during endurance exercise lasting longer than 70 minutes, consumed in the form of a carbohydrate-electrolyte beverage with a 6–8% concentration, 6–12 fluid ounces every 10–15 minutes throughout the exercise session. Nevertheless, while these existing protocols are based on robust research and are generally effective, they often recommend specific ranges of carbohydrate intake at fixed time intervals for different exercise durations, without accounting for individual differences and physiological responses. This oversight could potentially lead to over- or under-feeding, ultimately compromising athletic performance.

To determine the most appropriate amounts and timings for carbohydrate refueling during endurance exercise, measures of blood glucose, lactate levels, and heart rates may be beneficial, as they could objectively reflect the exercise-induced physiological responses of an individual [2,3]. However, monitoring glucose dynamics during endurance exercise is often limited by the traditional blood glucose measurement method, which requires frequent finger pricks and can be inconvenient to perform during exercise. Without real-time glucose data, tailoring personalized carbohydrate refueling strategies based on individual needs becomes challenging, which may increase the likelihood of experiencing suboptimal glycemic episodes, such as hypoglycemia, during training and competition [6].

Originally developed for diabetic patients, continuous glucose monitoring (CGM) technology enables real-time monitoring of interstitial glucose levels without the need for frequent finger pricks. It provides a potential solution for monitoring dynamic glucose fluctuations during physical activities [6]. This capability can be leveraged to assess glucose responses in individuals during endurance exercise, providing valuable insights for determining optimal carbohydrate refueling strategies [7–9]. While the application of CGM technology in sports is still in its early stages, some preliminary studies have documented its use in monitoring glucose dynamics during specific sporting events, such as ultra-marathons [10,11], cycling [12], and high-intensity interval training [13]. Overall, these studies suggest that CGMs can effectively determine the duration, magnitude, and frequency of interstitial glucose fluctuations, providing a unique opportunity for timely feedback for athletes to make real-time adjustments in response to undesirable interstitial glucose changes during competitive events and training sessions, potentially optimizing performance [7].

Despite the promising insights from these studies and the enthusiastic endorsements from athletes regarding the benefits of CGM technology [7], to the best of our knowledge, there is currently no well-established CGM-informed carbohydrate refueling protocol specifically designed for endurance athletes. Nonetheless, a position statement regarding glucose management for exercise using CGM in diabetic patients, endorsed by Juvenile Diabetes Research Foundation (JDRF) and supported by the American Diabetes Association (ADA), emphasizes the prevention of hypoglycemia during exercise for diabetic patients by utilizing both absolute glucose readings and trends provided by CGM devices to determine optimal carbohydrate replenishment timings [14]. Although the guideline may not be directly applicable to usage of CGM in healthy individuals and athletes, it leads to speculation of possible use of CGM in blood glucose informed nutritional strategies in such populations.

In this context, we designed the present exploratory study to compare the effectiveness of a traditional interval-based (TRAD) carbohydrate refueling protocol and a CGM-informed protocol designed by our research group. It is hypothesized that participants using the CGM-informed protocol would exhibit a more stable blood glucose trend and improved physiological responses during exercise when compared to those using the TRAD protocol during the same exercise trial. The outcomes may provide both theoretical and practical insights for sports professionals and athletes, enabling the integration of CGM technology into carbohydrate refueling strategies, which can ultimately lead to precise and effective nutrition approaches for enhancing athletic performance.

2. Methods

2.1. Study design

This study utilized a crossover, randomized controlled trial design. A total of 12 healthy, non-diabetic, recreationally active men (aged 18–45 years) with a general interest in endurance exercise were recruited via advertisements at universities, partner institutions, community centers, and social media platforms. Recreationally active participants were selected to assess the general efficacy of a CGM-informed carbohydrate refueling protocol, reflecting the target demographic of commercially available CGM products marketed to amateur athletes and fitness enthusiasts for performance optimization, and aligning with sports nutrition guidelines applicable to both professional and amateur populations. Baseline physical activity levels, assessed using the International Physical Activity Questionnaire (IPAQ) [15], confirmed all participants met the World Health Organization’s guidelines (150–300 minutes of weekly moderate-intensity physical activity) [16]. Additionally, participants were required to have engaged in endurance exercise lasting 90 minutes or more and be free from diagnosed cardiovascular, musculoskeletal, or metabolic diseases in the past six months to ensure they could safely complete the study protocol. All participants participated in a baseline test to determine their anthropometries, baseline glucose responses and aerobic fitness levels, followed by two main trials conducted at a consistent time of day (starting between 8:30 am and 12:00 noon, ±1 hour) to minimize diurnal variation. A 5–7 day washout period was implemented after the baseline test (prior to the first main trial) and between the first and second main trials to ensure full recovery from muscle fatigue and minimize carryover effects. Additionally, participants were required to avoid all planned exercise and engage only in normal daily activities for at least 48 hours prior to their scheduled main trials. To minimize the confounding influence of baseline glycogen levels, participants were also asked to fill in a food log and consume similar diet for 48 hours prior to each of the main trials to standardize baseline conditions. The study protocol adhered to the Declaration of Helsinki and was approved by the Research Ethics Committee of the HKU SPACE (rec@hkuspace.hku.hk).

Given the lack of direct comparisons between CGM-informed and TRAD protocols in existing studies, the sample size for this proof-of-concept study was determined based on a review of relevant literature on CGM applications in endurance athletes, which typically involved 7 to 12 participants [10,17–19]. A target sample size of 12 participants was chosen, balancing the need for meaningful data with the practical challenges of recruiting and retaining participants for a demanding exercise study. To further contextualize the findings, post-hoc effect size analyses were conducted, providing insights into the practical significance of the observed results.

2.2. Baseline test

Participants were requested to fast for at least 8 hours prior to the baseline test. During this test, body weight and body composition of the participants were measured using a bioelectrical impedance analyzer (MC-780 MA, Tanita, Japan). Participants were assessed barefoot and with light clothing after being asked to void their bladders. Following this, they rested quietly for 20–30 minutes before undergoing an oral glucose tolerance test (OGTT). This test was conducted to establish baseline glucose metabolism and to identify any abnormalities that could influence exercise responses. Finger-prick glucose measurements were taken using a glucometer (Contour Plus, Bayer Healthcare, Germany) before the ingestion of 75 grams of pure glucose (dextrose) dissolved in 250 mL of distilled water, and again at 2 hours post-ingestion. A peak power output (PPO) test was performed using a cycle ergometer (Monark LC7TT, Monark, Sweden) to assess baseline aerobic fitness levels and to establish the appropriate exercise intensity for the main trials. After the baseline test, each participant was provided with a CGM device (FreeStyle Libre, Abbott, USA) and was instructed to install it at least 24 hours before their first scheduled main trial.

2.3. Peak power output (PPO) test

A ramp incremental exercise test performed on an ergometer bike (LC7TT, Monark, Sweden), as adapted from a previously published protocol was utilized [20]. The pedal rate was self-selected and maintained between 60–80 rpm. Initially, participants performed 3 minutes of baseline cycling at 40 W. Thereafter, the work rate was increased at a rate of 30 W per minute until reaching the limit of tolerance. The test was complete when the pedal rate fell below 55 rpm despite verbal encouragement. The power output achieved at the point of exhaustion was recorded as the peak power output.

2.4. Main trials and exercise protocol

The exercise protocol is a modified version of an existing cycling test protocol [21]. During the two main trials, participants engaged in a 75-minute continuous cycling at an intensity of 50% of PPO, as determined in baseline testing, under a fasted state (8 to 12 hours). Participants followed either the TRAD or the CGM-informed carbohydrate refueling protocol in a randomized order. The randomization sequence was determined using the Excel randomization function on the day of the first main trial to ensure unbiased allocation. The two trials were separated by a 5–7 day-washout period. During each 75-minute cycling trial, glucose trends and various performance-related physiological metrics were measured to compare the effects of the two protocols. Finger-prick blood glucose and blood lactate were measured with portable analyzers (Contour Plus glucometer, Bayer Healthcare, Germany; Lactate Plus, Nova Biomedical, Waltham, MA, USA) before, in the middle and at the end of the cycling trial. The CGM glucose readings were captured before and after, and every five minutes during the cycling trial. The Borg’s 6–20 Rating of Perceived Exertion (RPE) [22] and heart rate were recorded every 15 minutes during the cycling trial (H10 Sensor, Polar Electro, Finland).

2.5. Traditional interval-based carbohydrates refueling protocol

The TRAD protocol followed the aforementioned guidelines from ACSM [5] and ISSN [4]. A serving of standard carbohydrate solution consisted of 208 mL (7 fluid ounces) of zero calorie sports drink (Aquarius Zero) combined with 16.7 g of pure maltodextrin (Nutricia PolyCal) (16 g, 7.7% carbohydrates) was given to the participants at 15, 30, 45, 60 minutes after beginning of the cycling trial (totaling 51.2 g carbohydrates and 665.6 mL fluid (22.5 fluid ounces) per hour).

2.6. CGM-informed carbohydrates refueling protocol

In the CGM-informed carbohydrates refueling protocol, glucose trends were used to determine the timings of carbohydrates replenishment. A serving of standard carbohydrates solution (same as above) was given when a falling blood glucose trend was identified, which was defined by one of the following four scenarios indicated by the CGM device: two consecutive glucose reductions of 0.1–0.2 mmol/L; or a single glucose reduction greater than 0.2 mmol/L; or a “Falling” arrow followed by a single glucose drop of 0.1–0.2 mmol/L; or a single “Falling quickly” arrow (further details on trend arrows can be referred to manufacture’s guidelines [23]). These scenarios help guide timely carbohydrate intake to prevent hypoglycemia during endurance activities. If no definitive falling blood glucose trend was identified for 15 minutes, a serving of placebo (208 mL Aquarius Zero alone) was given instead to ensure adequate hydration (i.e. similar fluid intake in both trials). The time interval between servings of standard carbohydrate solution was set to be a minimum of 15 minutes to allow sufficient time for the absorbed glucose to be reflected in the blood glucose readings.

2.7. Data analysis

All data were analyzed using the Statistical Package for the Social Sciences (SPSS), version 29.0.1 (IBM Corp., Armonk, NY, USA). The normality of the data was assessed using the Shapiro-Wilk test. Outcomes of physiological strain (e.g. HR, blood glucose, blood lactate) and perceived psychological strain (e.g. RPE) were assessed using two-way repeated measures ANOVA (refueling protocol × time). Refueling protocol and time were treated as the main effects in the statistical model. Post hoc comparisons were conducted with Bonferroni correction when a significant interaction between wearing refueling protocol and time was observed. All interstitial glucose data recorded during each main trial were converted into the total area under the curve (tAUC) and incremental area under the curve (iAUC) using a trapezoidal method and by using the baseline from timepoint 0 to calculate iAUC using GraphPad Prism 10.2.1 (GraphPad Software, San Diego, CA, USA). The coefficient of variation (CV%) was calculated to measure the relative variability of glucose levels by dividing the standard deviation by the mean glucose concentration. Statistical significance was set at a p value of < 0.05. Partial eta squared (η2p) and Cohen’s d effect size were calculated to indicate the magnitude of differences among experimental conditions where appropriate [24].

3. Results

Twelve healthy, non-diabetic males were recruited for this study (age = 32.3 ± 6.3 years, body mass = 63.5 ± 5.8 kg, body mass index = 21.7 ± 1.5 kg/m2, body fat percentage = 13.9 ± 4.2%, and PPO = 247.3 ± 39.9W). Participants reported an average of 298.8 ± 104.5 minutes of moderate-to-vigorous intensity physical activity per week. All participants successfully completed the two main trials with no adverse events reported. Fasting blood glucose (5.27 ± 0.29 mmol/L) and 2-hour blood glucose (5.98 ± 1.15 mmol/L) levels from the OGTT confirmed that all participants were below the diagnostic criteria for diabetes mellitus (i.e. fasting < 7.0 mmol/L and 2-hour < 11.1 mmol/L) [25].

3.1. Baseline carbohydrates intake

The average carbohydrate intakes of participants during the two days preceding the main trial were compared for each carbohydrate refueling protocol. Participants consumed an average of 4.69 ± 0.73 g of carbohydrates per kilogram of body weight per day (range = 3.74–6.01 g/kg/day) and 4.91 ± 0.97 g/kg/day (range = 3.88–7.51 g/kg/day) prior to the CGM-informed and TRAD trial respectively. The difference between trials was insignificant (p = 0.236), which suggested similar baseline glycogen storage that was unlikely to contribute to performance differences between the two main trials.

3.2. Carbohydrates supplementation amounts and timings

Table 1 compares the amount and timing of carbohydrate supplementation between the two carbohydrate refueling trials.

Table 1.

Comparison of the quantity and timing of carbohydrate supplementation between the two carbohydrate refueling protocols.

  CGM-informed
(N = 12)		 TRAD
(N = 12)
No. of times of standard carbohydrates solution supplementation 1.42 ± 0.79 (0–3) 4
Time of the initial supplementation (min) 31.8 ± 21.0 (15–65) 15
Time of the last supplementation (min) 50.5 ± 17.8 (15–70) 60
Average carbohydrate intake per hour (g) 18.1 ± 10.1 (0–38.4) 51.2
Open in a new tab

Data are presented as Mean ± SD (range).

CGM: Continuous glucose monitoring; TRAD: Traditional interval-based.

Regarding the administration frequency of the standard carbohydrate solution, participants consumed the solution on an average of 1.42 ± 0.79 times (range = 0–3 times) during the CGM-informed trial, thereby consuming an average of 18.1 ± 10.1 g of carbohydrates per hour (range = 0–38.4). In contrast, participants consumed the standard carbohydrates solution for 4 times at regular intervals during the TRAD trial, summing to an hourly intake of 51.2 g carbohydrates.

In terms of the timing of consuming standard carbohydrates solution, the average time of the initial carbohydrate supplementation was 31.8 ± 21.0 minutes (range = 15–65 minutes) after the start of the cycling trial in the CGM-informed trial, while participants received their first carbohydrate supplementation at 15 minutes after the start of the cycling trial in the TRAD trial. Additionally, participants consumed their last carbohydrate supplementation at 50.5 ± 17.8 minutes (range = 15–70 minutes) after the start of the cycling trial when following the CGM-informed protocol, while they received their last carbohydrate supplementation at 60 minutes after the start of the cycling trial during the TRAD protocol.

3.3. Assessments of blood glucose responses

The blood glucose responses comparing the two carbohydrate refueling protocols during the 75-minute exercise task are shown in Figure 1 and Supplementary Table1. The CV% in the CGM-informed condition (8.82 ± 4.68 %) was significantly lower than the TRAD condition (11.95 ± 3.41 %, Cohen’s d = 0.76, p = 0.013), indicating more stable glucose levels. The mean 75-min interstitial glucose tAUC responses in the TRAD condition (448.6 ± 38.2 mmol·min·L−1) was significantly higher than the CGM-informed condition (423.4 ± 39.1 mmol·min·L−1, Cohen’s d = 0.65, p = 0.008, Figure 2). In addition, the mean 75-min interstitial glucose iAUC responses in the TRAD condition (59.8 ± 21.9 mmol·min·L−1) was significantly higher than the CGM-informed condition (45.0 ± 28.5 mmol·min·L−1, Cohen’s d = 0.65, p = 0.018). Two-way repeated measures ANOVA revealed a significant refueling protocol × time interaction effect for interstitial glucose levels (p = 0.032, F2.903 = 3.169, η2p = 12.09). Post-hoc pairwise comparisons indicated that glucose levels in the TRAD condition were significantly higher than in the CGM-informed condition at the 60th, 65th, 70th, and 75th minutes (Figure 1 and Supplementary Table1).

Figure 1.

Figure 1.

Open in a new tab

Glucose trends between the two carbohydrate refueling protocols.

*p < 0.05

Figure 2.

Figure 2.

Open in a new tab

Mean 75-minute interstitial glucose tAUC responses.

3.4. Assessments of performance-related physiological responses

Heart rate, RPE, and blood lactate responses of the two carbohydrate refueling protocols during the 75-minute exercise task are presented in Supplementary Table1. The mean heart rate was 135.1 ± 14.0 bpm for the TRAD protocol and 133.2 ± 12.7 bpm for the CGM-informed protocol. Mean RPE was 13.4 ± 1.9 for the TRAD protocol and 12.9 ± 1.9 for the CGM-informed protocol. Blood lactate levels averaged 2.6 ± 1.4 mmol/L in the TRAD condition and 2.3 ± 0.6 mmol/L in the CGM-informed condition. Two-way repeated measures ANOVA revealed significant main effects of time for heart rate, RPE, and blood lactate levels (all p < 0.01), indicating general temporal changes in these metrics over the 75-minute cycling trial. However, no significant main effects of refueling protocol or protocol × time interactions were found (all p > 0.05). Overall, these findings indicate that heart rate, RPE, and blood lactate levels remained comparable between the two carbohydrate refueling protocols during the exercise task.

4. Discussion

To the best of our knowledge, this is the first study to directly compare a CGM-informed carbohydrate refueling protocol with a TRAD protocol based on existing sports nutrition guidelines. Contrary to our experimental hypothesis, the physiological outcomes – including heart rate, blood lactate, and RPE – which were employed to assess participants’ subjective and objective fatigue levels, did not show significant differences between the two carbohydrate refueling protocols. These findings raise important questions about the effectiveness of personalized carbohydrate strategies in influencing fatigue metrics during endurance exercise.

Several possible reasons may account for this lack of significant findings. Firstly, the duration of the exercise protocol (i.e. 75 minutes) and the intensity (50% peak power output) may not have been sufficient to elicit substantial differences in physiological responses. Previous research suggests that carbohydrate needs increase with greater duration and intensity [26]. The benefits of personalized carbohydrate strategies may become more pronounced during longer or more intense endurance efforts, where glycogen depletion and metabolic demands are heightened [26,27]. Additionally, the timing and quantity of carbohydrate intake in both protocols may have adequately met participants’ energy needs, potentially masking any differences in fatigue levels. The subtle variations in performance and fatigue levels between the protocols could have limited the sensitivity of the physiological measures employed (i.e. heart rate, lactate levels, and RPE) to detect significant differences. Furthermore, the exercise intensity for our endurance task was controlled and predetermined at 50% PPO. While this methodological approach enables a fair comparison of physiological and perceptual responses at a consistent intensity across the two carbohydrate refueling conditions, it may have limited the ability to assess changes in more direct performance metrics, such as mean power, completion time, or distance covered.

Despite the lack of positive findings with the above physiological measurements, our findings indicate that the CGM-informed protocol resulted in a lower yet more stable glucose trend during the 75-minute cycling trial, as evidenced by the CV%, iAUC and tAUC analyses. Conversely, the TRAD protocol, which adhered to established carbohydrate replenishment guidelines, demonstrated an increasing trend in blood glucose levels, particularly in the latter part of the trial. Current authoritative guidelines recommend the consumption of carbohydrates at a rate of approximately 30–60 grams per hour with a 6–8% carbohydrate-electrolyte solution every 10–15 minutes, during exercise lasting longer than an hour [4,5]. This recommendation is primarily based on the understanding that exogenous carbohydrate oxidation rates peak at 1–1.1 grams per minute, suggesting saturation of the sodium-glucose linked transporters (SGLT1) in the small intestine [28]. However, our results imply that rigid adherence to these guidelines (i.e. providing participants with 51.2 grams of carbohydrates per hour in the TRAD protocol at regular 15-minute intervals) may not adequately account for individual metabolic responses during exercise, as evidenced by the increasing glucose levels observed. In contrast, the CGM-informed protocol resulted in a mean carbohydrate intake of 18.1 ± 10.1 g/h (range 0–38.4 g/h), with the wide individual variation reflecting diverse glucose utilization patterns among participants. This variability suggests that CGM-informed protocol enables participants to precisely adjust their carbohydrate intake in response to real-time glucose trends, which may enhance metabolic control based on body’s actual needs, potentially benefiting athletic performance in endurance sports [8,29]. For instance, individuals prone to gastrointestinal distress from higher carbohydrate consumption may benefit from knowing that lower carbohydrate intake can elicit similar physiological responses, thereby improving tolerance and comfort during exercise. Nonetheless, blood glucose regulation is a complex process and previous studies utilizing CGM have consistently shown that athletes exhibit highly individualized glucose profiles, often spending significant time with hypo- and hyperglycemia states [8]. Future research should further investigate the practical implications of these observed differences in glucose responses between protocols. Specifically, it will be important to assess whether the stabilizing effect on blood glucose levels can positively impact overall performance in longer-duration endurance events where the importance of carbohydrate refueling will be amplified. Understanding these dynamics could lead to more tailored nutrition strategies that better support athletes’ metabolic needs during exercise.

Despite its potential applications in sports, the practicality of wearing CGM devices during exercise remains an important concern in real-world application. While all our participants in the study reported minimal discomfort from wearing the CGM sensors during their trial, frequent sensor detachment was notable. Four out of the twelve participants experienced sensor detachment and required replacement during the washout period after the first main trial. This suggests that the durability and adhesion of current CGM sensor technology may not be sufficient to withstand the mechanical stress, sweat, and movements of exercise, particularly in sports like swimming with high water exposure [7,30]. Moreover, a critical limitation of CGM technology is the time lag between interstitial and blood glucose measurements. While CGM can provide a general trend of glucose changes during exercise, the typical delay of 5 to 15 minutes may hinder the timely detection of glucose fluctuations [7,31]. This delay could impair the decision timing and effectiveness of CGM-informed refueling, especially in high-intensity intermittent exercise where glucose levels change rapidly, potentially leading to underfueling or suboptimal performance [32]. However, evidence from our own work in comparable exercise modalities suggests this limitation may be less pronounced during submaximal, continuous endurance exercise. We have previously demonstrated that despite a measurable time lag, the FreeStyle Libre FGM system maintains clinically acceptable accuracy (with 99.6–100% of readings in error grid zones A and B) during steady-state walking at 50% VO₂max [33]. Given that the present study employed a physiologically similar intensity (50% PPO cycling), where glucose fluctuations are more gradual, we believe the CGM-informed protocol provided reliable trend data for nutritional guidance. Nonetheless, cautious interpretation of real-time data and complementary strategies (e.g. preemptive fueling or hybrid monitoring with finger-prick tests) are warranted. These challenges underscore the need for further advancements in CGM sensor design, signal processing, and predictive algorithms to enhance reliability and accuracy during physical activity. Addressing these technical limitations is an important consideration as CGM becomes more widely adopted for metabolic health management and athletic performance optimization.

4.1. Implications for future studies

Our study contributes to the limited body of research investigating the comparative effects of CGM-informed and TRAD carbohydrate refueling protocols on endurance exercise responses. It offers new insights into the potential applications and practical concerns surrounding CGM technology in this context. However, it is also important to acknowledge the limitations of this exploratory study. First, the sample size was not determined by a priori calculation due to the absence of existing studies directly comparing CGM-informed and TRAD protocols. While this may limit the statistical power to detect difference in certain outcomes, we performed post-hoc effect size analyses to help interpret the practical significance of the findings and offer insights for future research. In addition, we were unable to perform subgroup analyses based on age groups due to the sample size constraint, but there is evidence indicating that age can influence CHO metabolism and glucose responses during endurance exercise [34,35], which supports the value of exploring such comparisons in future research with larger cohorts. Moreover, the 75-minute exercise protocol utilized in this investigation, while aligned with ACSM and ISSN guidelines for endurance exercise, was relatively short compared to some real-world endurance sports, such as marathons, triathlons, and ultramarathons, which often last for several hours. Future research should explore the effects over a broader range of exercise durations. Due to the recreational nature of our participants’ activity levels and lack of standardized training, we did not evaluate the effects of training experience or weekly volume on protocol responses. The use of recreationally active individuals may also limit generalizability to professional athletes, as training status could affect protocol outcomes. Future studies should compare training status (e.g. recreational vs. professional athletes) to evaluate its impact on CGM-informed versus traditional carbohydrate refueling protocol efficacy. Additionally, the specific physiological measures included used in this study may not necessarily reflect direct performance outcomes. Incorporating direct performance measures in future studies, such as time to completion or distance covered, may yield evidence regarding the practical implications of the CGM-informed carbohydrate refueling approach. Lastly, there is currently no universal CGM-informed refueling protocol specifically designed for endurance exercises. The CGM-informed protocol used in this study was developed based on practical experience, existing literature, and glucose-related metrics (e.g. falling trend) provided by CGM. Future studies should compare different CGM-informed protocols with varying glucose cutoff values to identify the most effective strategies for endurance performance.

5. Conclusion

This study provides valuable insights into the applications of a CGM-informed carbohydrates refueling protocol for endurance exercises. The physiological outcomes assessed in this study did not demonstrate statistically significant differences between the CGM-informed and TRAD protocols. However, the observed trends in blood glucose regulation suggest that the CGM-informed protocol may help maintain stable glucose levels, aligning carbohydrate intake with the participants’ actual metabolic needs during exercise. Further research is warranted, particularly with longer-duration exercise protocols and varied CGM-informed designs, to identify the optimal conditions under which these protocols can enhance athletic performance. By exploring these avenues, both researchers and practitioners can better understand the potential benefits of personalized carbohydrate strategies in endurance sports.

Supplementary Material

Supplemental Material
RSSN_A_2561670_SM2716.docx (19.6KB, docx)

Acknowledgments

We would like to thank all the participants in this study. We also thank the laboratory staffs for conducting the blinding for the solutions and randomization for the crossover order.

Funding Statement

This work was supported by the Chinese University of Hong Kong under United College Endowment Fund Research Grant & Lee Hysan Foundation Research Grant Schemes.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Authorship

EP and KL conceptualized the study design. KL collected the data, performed the data analysis and prepared the first draft. EP, JT, XZ, and WW revised the manuscript. EP supervised this study.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15502783.2025.2561670

References

  • 1.Hearris MA, Hammond KM, Fell JM, et al. Regulation of muscle glycogen metabolism during exercise: implications for endurance performance and training adaptations. Nutrients. 2018. Mar 2;10(3):298. doi: 10.3390/nu10030298 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Naderi A, Gobbi N, Ali A, et al. Carbohydrates and endurance exercise: a narrative review of a food first approach. Nutrients. 2023. Mar 11;15(6):1367. doi: 10.3390/nu15061367 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Vitale K, Getzin A.. Nutrition and supplement update for the endurance athlete: review and recommendations. Nutrients. 2019. Jun;11(6):1289. doi: 10.3390/nu11061289 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Kerksick CM, Arent S, Schoenfeld BJ, et al. International Society of Sports Nutrition position stand: nutrient timing. J Int Soc Sports Nutr. 2017;14(1):33. doi: 10.1186/s12970-017-0189-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Thomas DT, Erdman KA, Burke LM. American college of sports medicine joint position statement. Nutr And Athletic Perform Med Sci Sports Exerc. 2016. Mar;48(3):543–10. [DOI] [PubMed] [Google Scholar]
  • 6.Podlogar T, Wallis GA. New horizons in carbohydrate research and application for endurance athletes. Sports Med. 2022. Dec;52(Suppl 1):5–23. doi: 10.1007/s40279-022-01757-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Bowler AM, Whitfield J, Marshall L, et al. The use of continuous glucose monitors in sport: possible applications and considerations. Int J Sport Nutr Exerc Metab. 2023. Mar 1;33(2):121–132. doi: 10.1123/ijsnem.2022-0139 [DOI] [PubMed] [Google Scholar]
  • 8.Flockhart M, Larsen FJ. Continuous glucose monitoring in endurance athletes: interpretation and relevance of measurements for improving performance and Health. Sports Med. 2024. Feb;54(2):247–255. doi: 10.1007/s40279-023-01910-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Schubert-Olesen O, Kroger J, Siegmund T, et al. Continuous glucose monitoring and physical activity. Int J Environ Res Public Health. 2022. Sep 28;19(19):12296. doi: 10.3390/ijerph191912296 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Ishihara K, Uchiyama N, Kizaki S, et al. Application of continuous glucose monitoring for assessment of individual carbohydrate requirement during ultramarathon race. Nutrients. 2020. Apr 17;12(4):1121. doi: 10.3390/nu12041121 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Sengoku Y, Nakamura K, Ogata H, et al. Continuous glucose monitoring during a 100-km race: a case study in an elite ultramarathon runner. Int J Sports Physiol Perform. 2015. Jan;10(1):124–127. doi: 10.1123/ijspp.2013-0493 [DOI] [PubMed] [Google Scholar]
  • 12.Iscoe KE, Campbell JE, Jamnik V, et al. Efficacy of continuous real-time blood glucose monitoring during and after prolonged high-intensity cycling exercise: spinning with a continuous glucose monitoring system. Diabetes Technol Ther. 2006. Dec;8(6):627–635. doi: 10.1089/dia.2006.8.627 [DOI] [PubMed] [Google Scholar]
  • 13.Bally L, Zueger T, Pasi N, et al. Accuracy of continuous glucose monitoring during differing exercise conditions. Diabetes Res Clin Pract. 2016. Feb;112:1–5. doi: 10.1016/j.diabres.2015.11.012 [DOI] [PubMed] [Google Scholar]
  • 14.Moser O, Riddell MC, Eckstein ML, et al. Glucose management for exercise using continuous glucose monitoring (CGM) and intermittently scanned CGM (isCGM) systems in type 1 diabetes: position statement of the European Association for the Study of Diabetes (EASD) and of the International Society for Pediatric and Adolescent Diabetes (ISPAD) endorsed by JDRF and supported by the American Diabetes Association (ADA). Diabetologia. 2020. Dec;63(12):2501–2520. doi: 10.1007/s00125-020-05263-9 [DOI] [PubMed] [Google Scholar]
  • 15.Macfarlane DJ, Lee CC, Ho EY, et al. Reliability and validity of the Chinese version of IPAQ (short, last 7 days). J Sci Med Sport. 2007. Feb;10(1):45–51. doi: 10.1016/j.jsams.2006.05.003 [DOI] [PubMed] [Google Scholar]
  • 16.Bull FC, Al-Ansari SS, Biddle S, et al. World Health Organization 2020 guidelines on physical activity and sedentary behaviour. Br J Sports Med. 2020. Dec;54(24):1451–1462. doi: 10.1136/bjsports-2020-102955 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Bowler AM, Burke LM, Coffey VG, et al. Day-to-day glycemic variability using continuous glucose monitors in endurance athletes. J Diabetes Sci Technol. 2024. May 10. doi: 10.1177/19322968241250355 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Coates AM, Thompson KMA, Grigore MM, et al. Altered carbohydrate oxidation during exercise in overreached endurance athletes is applicable to training monitoring with continuous glucose monitors. Scand J Med Sci Spor. 2024. Jan;34(1). doi: 10.1111/sms.14551 [DOI] [PubMed] [Google Scholar]
  • 19.Kulawiec DG, Zhou T, Knopp JL, et al. Continuous glucose monitoring to measure metabolic impact and recovery in sub-elite endurance athletes. Biomed Signal Proces. 2021. Sep;70:103059. doi: 10.1016/j.bspc.2021.103059 [DOI] [Google Scholar]
  • 20.Zhang JQ, Dai ZH, Stephen HSW, et al. Acute effects of various doses of nitrate-rich beetroot juice on high-intensity interval exercise responses in women: a randomized, double-blinded, placebo-controlled, crossover trial. J Int Soc Sports Nutr. 2024. Dec 31;21(1). doi: 10.1080/15502783.2024.2334680 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Kilian Y, Wehmeier UF, Wahl P, et al. Acute response of circulating vascular regulating microRNAs during and after high-intensity and high-volume cycling in children. Front Physiol. 2016;7:92. doi: 10.3389/fphys.2016.00092 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Borg E, Kaijser L. A comparison between three rating scales for perceived exertion and two different work tests. Scand J Med Sci Spor. 2006. Feb;16(1):57–69. doi: 10.1111/j.1600-0838.2005.00448.x [DOI] [PubMed] [Google Scholar]
  • 23.Libre F. Trend arrows and scanning through the day. 2025. [cited 2025 Mar 20]. Available from: https://www.freestyle.abbott/om-en/discover-freestyle-libre/getting-started-with-freestyle-libre/trend-arrows-and-scanning-through-the-day.html?srsltid=AfmBOooUC7vy43QO5Y128kCjVouhoQxaCbVOG46dAKBHE4GsnlIOxX4H
  • 24.Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale, N.J: L. Erlbaum Associates; 1988. [Google Scholar]
  • 25.American Diabetes Association Professional Practice Committee . Classification and diagnosis of diabetes: standards of medical care in diabetes-2022. Diabetes Care. 2022 Jan 1;45(Suppl 1):S17–S38. doi: 10.2337/dc22-S002 [DOI] [PubMed] [Google Scholar]
  • 26.Jeukendrup A. A step towards personalized sports nutrition: carbohydrate intake during exercise. Sports Med. 2014. May;44(Suppl 1):S25–33. doi: 10.1007/s40279-014-0148-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Burke LM, Hawley JA, Wong SH, et al. Carbohydrates for training and competition. J Sports Sci. 2011;29(Suppl 1):S17–27. doi: 10.1080/02640414.2011.585473 [DOI] [PubMed] [Google Scholar]
  • 28.Jeukendrup AE, Jentjens R. Oxidation of carbohydrate feedings during prolonged exercise: current thoughts, guidelines and directions for future research. Sports Med. 2000. Jun;29(6):407–424. doi: 10.2165/00007256-200029060-00004 [DOI] [PubMed] [Google Scholar]
  • 29.Oishi A, Makita N, Kishi S, et al. Continuous glucose monitoring of a runner during five marathons. Sci Sport. 2018. Dec;33(6):370–374. doi: 10.1016/j.scispo.2018.05.001 [DOI] [Google Scholar]
  • 30.Hughes MS, Kingman RS, Hsu L, et al. Swimming with the Omnipod 5 automated insulin delivery system: connectivity in the water. Diabetes Care. 2023. Aug 1;46(8):e148–e150. doi: 10.2337/dc23-0470 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Siegmund T, Heinemann L, Kolassa R, et al. Discrepancies between blood glucose and interstitial glucose-technological artifacts or physiology: implications for selection of the appropriate therapeutic target. J Diabetes Sci Technol. 2017. Jul;11(4):766–772. doi: 10.1177/1932296817699637 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Li A, Riddell MC, Potashner D, et al. Time lag and accuracy of continuous glucose monitoring during high intensity interval training in adults with type 1 diabetes. Diabetes Technol Ther. 2019. May;21(5):286–294. doi: 10.1089/dia.2018.0387 [DOI] [PubMed] [Google Scholar]
  • 33.Zhang X, Sun F, Wongpipit W, et al. Accuracy of flash glucose monitoring during postprandial rest and different walking conditions in overweight or obese young adults. Front Physiol. 2021;12:732751. doi: 10.3389/fphys.2021.732751 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Clevenger CM, Parker Jones P, Tanaka H, et al. Decline in insulin action with age in endurance-trained humans. J Appl Physiol (1985). 2002. Dec;93(6):2105–2111. doi: 10.1152/japplphysiol.00315.2002 [DOI] [PubMed] [Google Scholar]
  • 35.Sial S, Coggan AR, Carroll R, et al. Fat and carbohydrate metabolism during exercise in elderly and young subjects. Am J Physiol. 1996. Dec;271(6):E983–9. doi: 10.1152/ajpendo.1996.271.6.E983 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Material
RSSN_A_2561670_SM2716.docx (19.6KB, docx)

Articles from Journal of the International Society of Sports Nutrition are provided here courtesy of Taylor & Francis

RESOURCES