Abstract
Background:
This descriptive observational study reports on continuous glucose monitoring (CGM) data, using a novel glucose biosensor (Abbott Libre Sense Glucose Sport Biosensor), during professional game play and during daily life in elite European football players.
Methods:
Eighteen healthy male elite football players (age: 27.5 ± 5.1 years; height 180.1 ± 7.2 cm, weight 74.2 ± 9.1 kg, UEFA Champions League club) participated, with a subset examined for a single game for active (n = 10) and reserve (n = 4) players. Group comparisons used unpaired t-tests or Wilcoxon rank-sum tests; within-group differences used repeated measures one-way analysis of variance or Friedman test. Descriptive statistics were summarized for 24-hour data for daytime (06:00
Results:
Higher mean CGM glucose was observed during-game in active compared with reserve players (159 ± 23 vs 133 ± 25 mg/dL, P = .09), with significantly higher time above range (TAR, 72.8 ± 32.02 vs 29.7 ± 37.9%, P = .04) and lower time in range (TIR, 26.7 ± 31.9 vs 70.3 ± 37.9%, P = .04). In the 90 minute pre- to 180 minute post-game period, TAR (57.3 ± 26.6% vs 16.1 ± 20.2%, P = .02) and mean iG (149 ± 19 vs 123 ± 14 mg/dL, P = .02) remained higher for active players. For all 18 players, TIR was 89.4 ± 11.7 and 91.6 ± 13.7%, TAR was 5.9 ± 6.7 and 2.9 ± 5.7%, and time below range was 4.5 ± 10.5 and 5.3 ± 13.2% for day and night, respectively.
Conclusions:
This observational study suggests that elite European footballers may have significant increases in glycemia, as measured by CGM, supporting the notion that mild hyperglycemia can occur during and after active competition in healthy and metabolically normal athletes, perhaps because of competition stress.
Introduction
The glucose profiles of competitive athletes who do not have diabetes have been profiled using continuous glucose monitoring (CGM) devices in several recent studies.1-3 While interstitial glucose concentrations, which are thought to be reasonably reflective of glucose levels in the bloodstream during exercise, 4 - 6 tend to remain in a relatively tight glycemic range (ie, 70-140 mg/dL; ~3.9-7.8 mmol/L) most of the time, observations outside of this so-called euglycemic range have been documented in individuals without diabetes. Specifically, in endurance athletes, glucose level have been reported to be within the normal glycemic range of 70 to 140 mg/dL (3.9-7.8 mmol/L) ~90% of the time, with level 1 hyperglycemia (141-180 mg/dL; ~7.9-10 mmol/L) observed ~5% of the time and with level 1 hypoglycemia (54-69 mg/dL; ~3.0-3.8 mmol/L) occurring ~5% of the time.2,7,8
Performance during team sports, such as European football (or American Soccer), is characterized by alternating periods of very high-intensity activity and moderate-to-light activity (or rest), which is dependent on a combination of anaerobic and aerobic energy systems. 9 During competitive games, professional players run at extremely high speeds (>19.8 km/h), jump, head the ball, pass, shoot, accelerate, decelerate, 10 but also must have a high endurance capacity as they often cover distances of 10 to 13 km per game, 11 with an estimated 90% of the time at a low-to-moderate intensity. 12 The nutritional intake of such players is mainly via high and varying amounts of carbohydrate intake (3-12 g × kg−1 body mass per day), but with little to no carbohydrate consumed during competitive game play. 12 How these types of nutritional strategies, and the training and competition events associated with the sport of European football, affect CGM glucose is unknown. Moreover, it is unclear whether differences may exist in CGM values between active (ie, on field) and nonactive (ie, reserve or bench) players, resulting perhaps from physical and/or emotional demands of on-field play.
This descriptive study aims to elucidate the patterns of CGM during an elite team’s football game and during “normal” day-to-day living. Furthermore, it seeks to characterize daily CGM trends of healthy elite footballers throughout the sensor wear time.
Methods
Eighteen (n = 18) elite-level male professional football players without diabetes (age: 27.5 ± 5.1 years; height 180.1 ± 7.2 cm, weight 74.2 ± 9.1 kg) playing for the same club (Union of European Football Associations [UEFA] Champions League) volunteered to participate in the study. Each player wore a new-generation CGM (Libre Sense Glucose Sport Biosensor, Abbott Diabetes Care, Alameda, California) for up to 14 days in real-life situations, which included: multiple official games at national and international levels, training, traveling, and so on. This analysis focused on daily trends throughout the sensor wear time, including the percent (%) time in the so-called “euglycemic range” (70-140 mg/dL; 3.9-7.8 mmol/L), % time above range (TAR; >140 mg/dL; >7.8 mmol/L) and time below range (TBR; <70 mg/dL; <3.9 mmol/L), in a subset of players (n = 14) during a single competitive game event that included both active and nonactive players. The glucose biosensor used measures interstitial glucose within a 3.0 to 11.1 mmol/L (55-200 mg/dL) minimum to maximum range and can report data every minute using an integrated surface recorder/transmitter that communicates in real time to a smartphone application (Supersapiens Inc., TT1 Products, Atlanta, Georgia) via Bluetooth technology. If the smartphone is out of range, such as during game play, CGM data are stored at a 15-minute sampling frequency for up to 8 hours. Running performance game data were measured and collected with by Global Positioning System GPS technology (S7 Vector, Catapult, Melbourne, Australia) with a sampling frequency of 10 Hz. During the game, players wore GPS vests, which included a placed GPS unit that was turned on 15 minutes before the start of the game.
The Supersapiens App User Agreement and Privacy Policy accepted by the user states that anonymized data can be used for public and/or research purposes.
Statistics
Footballers were categorized into two groups: (1) active players who participated in the game for at least 12 minutes (n = 10) and (2) reserve players who played for less than 12 minutes (n = 4). Sensor glucose data were summarized for each player and analyzed using GraphPad Prism (Version 10.4.1).
All data were assessed for normality using the Shapiro-Wilk test. Glycemic outcomes included mean CGM glucose, maximum CGM values, minimum CGM values, glucose standard deviation (SD), glucose coefficient of variation (CV), % time in the normal glycemic range (TIR: 70-140 mg/dL; 3.9-7.8 mmol/L), % TBR (<70 mg/dL; <3.9 mmol/L), and % TAR (>140 mg/dL; >7.8 mmol/L).
For primary outcomes, CGM metrics were analyzed across four-time segments based on each player’s relative game start time: (1) pre-game (−90 to −1 minutes), (2) during game (0 to 105 minutes), (3) post-game (106 to 180 minutes), and (4) overall (−90 to 180 minutes). Metrics were compared between players and reserves using either an unpaired t-test (parametric) or Wilcoxon rank-sum test (nonparametric). Within-group differences across time segments were assessed using a repeated measures one-way analysis of variance (for normally distributed data) or a Friedman test (for non-normally distributed data), followed by Dunn’s multiple comparison test.
For secondary outcomes, glycemic metrics were summarized over the entire 14-day wear period for all team members (n = 18), stratified by daytime (06:00
Results
Game Analysis—Performance
For the ten active game players, average game playing time, distance covered, and maximum velocity were 62.67 ± 32.69 minutes, 6820.27 ± 3462.74 m, 31.47 ± 2.23 km/h, respectively. Grouped mean number of accelerations >21 km/h was 23 ± 14 and the distance covered at high intensity (21-24 km/h) was 193 ± 121 m during the game play. Grouped mean number of total sprints >24 km/h was 10 ± 7 and the distance covered at sprinting intensity (>24 km/h) averaged 211 ± 160 m. Players’ mean number of high-intensity accelerations (>3 m/s2) and decelerations (>−3 m/s2) were 44 ± 23 and 46 ± 23, respectively.
Game Analysis—CGM Glucose
The CGM time series for the players involved in the game (both those who took part on the game and those who sat on the bench) is reported in Figure 1 with in-game summary statistics reported in Table 1. In the pre-game period, CGM values were relatively stable, as measured by mean and SD of glucose concentrations, and were similar between groups (P = .06 and P = .82, respectively). During the game, CGM levels tended to rise in both groups, but with higher mean values observed in the active players (159 ± 23 vs 133 ± 25 mg/dL, P = .09). The CGM tracings were also higher for active players compared with reserve by the end of game play and during the 70-minute recovery period, with higher peak CGM values observed, but with similar minimum CGM levels, SD and CV scores also observed. Over the entire pre-, during, and post-game play, active players had higher mean CGM values than reserve players (149 ± 19 vs 123 ± 14 mg/dL; P = .02) and a higher percent time >140 mg/dL (TAR: 57.3 ± 26.6 vs 16.1 ± 20.2 %, P = .02). The number of active playing minutes correlated significantly with the CGM % time >140 mg/dL during match play (R 2 = .40; P = .02).
Sensor CGM glucose levels of active (blue) and reserve (pink) players before (−90 to 0 minutes), during (0 to 105 minutes), and after (105 to 180 minutes) an official game. Vertical dotted lines delineate game play. Solid lines represent the medians, and shaded areas indicate the IQR for each group. Sensor reads from 55 to 200 mg/dL.
Glycemic Outcomes From Game Data for the Professional Football Players Grouped by Their Participation in the Game (Active, n = 10; Reserve, n = 4) and Stratified by Time Period.
Abbreviations: Max CGM glucose, maximum CGM glucose level; Min CGM glucose, minimum CGM glucose level; SD, standard deviation of glucose; CV, coefficient of variation of glucose; TIR, time in range 70 to 140 mg/dL; TBR, time below range <70 mg/dL; TAR, time above range >140 mg/dL.
Bold type face indicates significance at P<0.05.
Significantly different within group compared with pre-game data.
Non-normally distributed data.
Significantly different within group compared with post-game data.
Significantly different within group compared with during-game data.
CGM 24-Hour Data
Summary statistics for all 18 members of the team for 24-hour CGM values are shown in Table 2. Day and night TIR were 89.4 ± 11.7% and 91.6 ± 13.7%, respectively. Day and night TAR were 5.9 ± 6.7% and 2.9 ± 5.7%, whereas TBR were 4.5 ± 10% and 5.3 ± 13.2%, respectively.
Summary Statistics for the CGM Glucose Values in all Professional Football Players (N = 18) During Their Sensor Wear Time Period.
Abbreviations: SD, standard deviation of glucose; CV, coefficient of variation of glucose; TITR, time in tight range 70 to 140 mg/dL; TBR, time below range <70 mg/dL; TAR, time above range >140 mg/dL.
Discussion
This study sought to characterize the patterns of CGM during an elite team’s football training and game play as well as everyday life by means of a novel sports biosensor and CGM smartphone application. Overall, our findings indicate that CGM levels were relatively stable in the pre-game period and similar between active and reserve players. However, during game play, a more pronounced rise in CGM values were seen in the active players compared with their reserve counterparts, with most active players developing overt hyperglycemia (i.e., 72.8% of all CGM values >140 mg/dL; >7.8 mmol/L) by the end of match play. Moreover, active players spent only 37% of the time within the so-called “normal glycemic range” with active players having >60% of all CGM values in the mild-to-moderate hyperglycemic range in the post-game period. These findings support the notion that significant elevations in CGM values may be expected, to values that are deemed indicative of hyperglycemia in settings of diabetes, in professional athletes who are using CGM.
The use of CGM in athletes without diabetes is becoming increasingly popular.1,8,13 Athletes who do not have diabetes may be interested in observing if their glucose levels are “optimal” or “normal” for training and competition. They may also use CGM to learn how their body responds to various fueling strategies. Unfortunately, only a handful of studies have examined how different types of competitive activities might influence glucose levels as measured by CGM in athletes without diabetes and some issues of CGM accuracy have been raised for this target market.4,14,15 Nonetheless, others report that CGM can be reflective of whole blood glucose levels during exercise and that some forms of intense exercise may cause glucose to rise even when carbohydrate feeding does not occur.4,6 The increases in glycemia with some intense activities appear related, at least in part, to elevations in circulation catecholamines that increase hepatic glucose production beyond what can be taken up by contracting skeletal muscles.16-19 In line with these earlier in laboratory studies, this observational study using CGM shows that professional-level football players spent considerable time during and after a competitive match with elevated glucose, as measured by CGM, particularly if they were active on-field players during the match. Similar to previous studies that relied on infrequent whole blood glucose measurements from indwelling catheters,20,21 transient decreases in glucose levels measured by means of the CGM were also visible after the game, but with little to no hypoglycemia observed.
It is currently unclear whether the apparent physiologic rise of in-exercise glucose levels serves as a competitive advantage for active and inactive players. However, as the brain predominantly uses glucose from circulation (and ultimately the interstitial fluid) as an energy source, and elevated blood glucose concentration has been associated with an overall improvement in skill performance in football 22 as well as improved decision-making and successful execution, 23 it is possible that mild-to-moderate hyperglycemia is somehow beneficial during such a competitive game. In line with this possibility, it was previously reported that a carbohydrate-electrolyte gel ingested before an extra time period raised blood glucose concentrations and improved dribbling performance during the extra time period of simulated football gameplay. 24 Although detailed nutritional data are unavailable for this study, the athletes generally consumed their last pre-game meal ~3.5 hours before the game and typically did not supplement with additional carbohydrates during the game. When players were interviewed, most only reported consuming water, and indicated that they refrained from consuming sports drinks or energy gels during the game. However, these claims could not be verified by the investigators.
The CGM levels of elite football players in this study appear to show exaggerated hyperglycemia during exercise relative to reference populations of healthy adults 25 or recreational athletes. 7 In line with our findings, one recent study by Flockhart and colleagues 26 reported hyperglycemia in a cohort of elite endurance athletes, particularly when intense exercise occurred. Moreover, the relationship between the number of active playing minutes and % TAR in our study reinforces the relationship between high exercise intensity (as shown by in-game GPS data) and glucose fluctuations. These findings emphasize the importance of CGM in understanding real-time metabolic responses in high-performance athletes.
Post-game glucose levels remained high for the active players group. Similar findings were observed, although in the endurance-setting, by Parent et al, 27 in a study with endurance athletes that participated in a 156-km ultra trail race. These researchers speculated that muscle damage, defined by high circulating levels of creatine kinase and lactate dehydrogenase during the recovery period, could promote postexercise insulin resistance and/or hyperglycemia that resolves after several hours of recovery. 27
During the day, the elite-level European football team in our study spent ~90% of their time with CGM values in the normal glycemic range (70-140 mg/dL), which is comparable to other studies that included elite-level athletes: 91% in para cyclists, 8 93% in a professional female cycling team, 13 and 89% in well-trained male cyclists 26 but also somewhat lower TIR than what has been reported a healthy less active population (ie, 96% TIR). 28 This difference might be caused by the addition of 3 to 4 hours of regular daily exercise, higher energy expenditure, and assumingly a higher daily carbohydrate intake. In fact, similar to what is reported by Weijer et al, 8 the cohort in this study had ~6% TAR during each 24-hour day (except for the match play day). While it is true that elite endurance athletes tend to consume more carbohydrates during training and competition than the general public, 12 TAR is typically low in trained athletes as they have high skeletal muscle insulin sensitivity and high rates of glucose disposal even when muscle lipid levels are elevated. 29 This study has several limitations. Due to practical constraints, we were unable to include more athletes in the game analysis, which would have allowed for a more robust comparison between active and reserve players. Access to players and competitive matches was inherently restricted, and the use of monitoring devices during Champions League fixtures required special permissions, further limiting the number of games and players available for inclusion. In addition, the balance between active and reserve players was dictated entirely by in-game dynamics, specifically the tactical decisions of the head coach regarding substitutions. However, it is important to emphasize that this study was conducted in collaboration with one of the highest-ranked teams in the UEFA Champions League, a multiple-time winner of the competition. In addition, we cannot entirely rule out the influence of diet on post-game glycemic metrics, particularly given that athletes are generally advised to consume carbohydrates as soon as possible after competition. Expanding this research to other professional football teams with better nutritional monitoring would provide further insights into glucose dynamics in elite athletes participating in intermittent sports.
Another potential advancement of this line of research would be the collection of A1c measurements, which would allow assessment of the longer-term impact of the observed glucose levels and help clarify their potential clinical significance. Future studies should therefore integrate CGM data with HbA1c levels and other metabolic biomarkers to better contextualize acute glucose fluctuations within the broader framework of long-term glucose regulation in elite athletes.
Conclusion
This study is the first to provide exploratory data on in-game and 24-hour CGM glucose levels in elite European football players. While the clinical significance of the observed CGM values, suggesting hyperglycemia exposure, is uncertain in athletes living without diabetes, this type of data provides preliminary insights into real-world glucose dynamics in elite athletes and lay the groundwork for future studies that integrate CGM with performance outcomes and various clinical assessments of metabolic health. Contrary to the assumption that moderate elevation of glucose levels is absent in healthy individuals during exercise, we observed episodes of hyperglycemia in these athletes, particularly those with higher active playing time. The 24-hour data revealed a time-in-range comparable to previous studies on elite athletes but distinct from data on the general healthy population. As this difference could be specific to elite-level athletes, future research should focus on understanding optimal glucose responses to training and refining nutrition strategies tailored to this group.
Practical Implications
Continuous glucose monitoring may serve as a valuable tool for assessing the metabolic demands of an elite football game by capturing real-time fluctuations in interstitial glucose levels.
The findings suggest that glucose levels can rise significantly during and after a highly competitive game, such as European football. Tailored post-game nutrition and recovery protocols, including carbohydrate replenishment and muscle recovery strategies, may be necessary.
Elite athletes might exhibit transient hyperglycemia associated with competition and further studies should explore sport-specific glucose patterns in high-performance setting.
Footnotes
Acknowledgements
The authors express their gratitude to the players who participated in the study for their dedication and commitment to the data collection process. Special thanks are also extended to the team’s medical and performance staff for their invaluable support and contributions throughout the project.
Abbreviations
CGM, continuous glucose monitoring; CV, coefficient of variation; GPS, global positioning system; TAR, time above range (>140 mg/dL or >7.8 mmol/L); TBR, time below range (<70 mg/dL or <3.9 mmol/L); TIR, time in range (70-140 mg/dL or 3.9-7.8 mmol/L); UEFA, Union of European Football Associations.
Author Contributions
KS: project administration, data curation, investigation, methodology, validation, writing original draft, writing—review and editing, supervision; AZ: data curation, formal analysis, investigation, validation, visualization, writing—review and editing; NM: project administration, methodology, investigation, writing—review and editing; DL: methodology, writing—review and editing; LVT: data curation, formal analysis, validation, visualization, writing—review and editing; MCR: conceptualization, investigation, methodology, supervision, validation, writing—review and editing; HCZ: conceptualization, investigation, methodology, supervision, validation, writing—review and editing. All authors have read and agreed to the published version of the manuscript.
Declaration of Conflicting Interests
The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Authors Kristina Skroce, Andrea Zignoli, and David J. Lipman served as consultants for Supersapiens Inc. (TT1 Products, Atlanta, GA, USA) during the data collection period. Michael C. Riddell has served as a scientific advisor for Supersapiens Inc. and is currently a scientific advisor to another CGM company (Dexcom Inc). Howard C. Zisser was an employee of Supersapiens Inc. during the data collection period. The remaining authors declare no conflicts of interest.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research received no external funding. Supersapiens Inc. provided the CGM supplies that were necessary to conduct the study.
ORCID iDs
Data Availability Statement
Aggregated data might be available after reasonable and kind request has been sent via email to the corresponding author.