Eight professional male (age: 28.0 ± 5.5 years, body mass:86.4 ± 4.8 kg, height: 1.8 ± 0.1 m) outfield footballers competing in the third tier of English football during the 2023/24 season volunteered for the study. Participants had no history of diabetes and provided written informed consent. Ethical approval was granted by Nottingham Trent University Ethics Committee.

Participants completed a 14-day control (112 total player days) period, during which they reported dietary intake. This was followed by a 28-day observation period (188 total player days), whereby participants wore the CGM (Abbot Libre Sense Glucose Sport Biosensor, United States) and continued to report dietary intake. Two participants withdrew from the study during the observation period (Figure 1).

CGM are reported to have a lag time of 1.2 - 7.0 minutes and >90% of sensor readings are within 20% of blood glucose concentrations 24. The device has an upper limit of 11.2 mmol/L. The device was inserted into the back of the upper arm by the researcher and replaced after 14 days, as per manufacturer instructions 25. Specialist manufacturer-supplied performance patches, made of a silicone protection ring and 4-way stretch kinesiology tape, were placed onto the device for protection. The CGM was paired with the user’s smartphone application 26 via Bluetooth for minute-by-minute data collection. If the device was disconnected from Bluetooth (e.g., due to distance between participants and their smartphone), readings were taken every 15 minutes and stored by the device for 8 hours until it was re-synchronised via Bluetooth. Post-training, at half-time and post-match, participants were instructed to re-synchronise the CGM device with their smartphone via Bluetooth.

Participants reported dietary intake using a Snap and Send method, aimed to reduce reporting bias 27. Participants took photos of all food and drink consumed alongside an object of known size (massage ball, 8cm diameter) at 45- and 90- degree angles. Participants received individual demonstrations regarding the dietary reporting method with common troubleshooting issues discussed, e.g., complex dishes with multiple ingredients to be listed alongside the photograph. Images were sent via an instant messenger service (Whatsapp, Meta Platforms Inc.) to the primary researcher. If participants had not sent a photo through within 2 hours of their usual predicted mealtime, the researcher reminded participants via a text message e.g., breakfast reminder was sent at 10:30 in anticipation that a meal would have been consumed around 08:30. If participants failed to photograph their meal, dietary recall information was provided with portion sizes estimated. The primary researcher was present when meals were consumed at the training centre and assisted with dietary reporting by photographing and documenting supplement use. Participants reported all supplement use as part of their food diary, examples include caffeine gum, beta-alanine, creatine, and electrolytes. If supplements were provided for the participants by the club, dosage was confirmed with the team’s nutritionist. For supplements taken at home, participants reported brand, weighed measure or dosage. On Day 2 of the observation period, the potential impact of tobacco free nicotine pouches, alongside tobacco containing Snus, on glucose concentration was observed and, therefore, nicotine pouch/Snus use was reported on Day 2 onwards. Nicotine pouch/Snus administration times and corresponding CGM value were time aligned based on the known delay (2.4 ± 4.6 minutes) with CGM readings 24.

The feasibility of CGM was assessed based on the percentage of completed training and match sessions during which the device remained in situ with data available for analysis. Additionally, any adverse effects reported by participants, particularly discomfort, were recorded.

Overnight fasted glucose concentrations were determined by calculating the mean of overnight glucose concentrations from 00:00-06:00 during Week 1 of the observation period. This mean was used as a baseline measure. Glucose concentrations were assessed three hours pre-match, first half, half-time, second half, three hours post-match, and for training sessions, for each participant. Mean concentrations are reported during each period. During training and match play, the Bluetooth connection to CGM was lost, due to the distance between the monitor and the player’s smartphone, a singular 5-minute mean glucose concentrations were recorded. As such, first and second half match play was restricted to three measurements in 15-minute intervals. Percentage and absolute change from overnight fasted interstitial glucose concentrations were calculated for all match-play and training session measurements.

A Shaprio-Wilk’s test was conducted to assess normality in distribution of data with all data confirmed as normally disturbed. Paired T-tests, with a significance level of p < 0.05, and Bonferroni correction were used to assess the difference between mean glucose concentrations during training sessions and match play (inclusive of first half, half time and second half), overnight fasted glucose concentrations and training, and overnight fasted glucose concentrations and match-play. A paired T-test was also used to assess differences in dietary intake between the control and observation periods.

Dietary intake was analysed by a Sports and Exercise Nutrition registered (SENr) dietitian, using dietary analysis software (Nutritics, Version 5). Dietary intake was quantified as energy (kilocalories), carbohydrates (grams) and protein (grams) in absolute units and relative to body mass. Mean energy and macronutrient intake between the control and observation period were compared to establish any significant changes for the duration of the study using a paired T-test, following normality testing (Nutritics, Version 5).

CGM were used in 148 training or match events, and needle displacement occurred on three occasions, all during training, thus the CGM produced data for 98% of training and match play events without displacement. Six participants regularly participated in training and match play. Four of the six participants sustained short-term muscular strain injuries, which resulted in a total of 9 player days of modified training i.e ‘off feet’ indoor gym work. Two participants were injured prior to the study start date and did not return to training or match play during the study period, participating only in rehabilitation sessions. However, these participants continued to wear the CGM and record dietary intake throughout their rehabilitation.

Carbohydrate intake did not differ between the control and observation periods (control 3.0 ± 0.6 g/kg/day, observation period 3.2 ± 0.7 g/kg/day) (t = 1.595, p = 0.077). There was no difference in energy and protein intake between the control and observation periods (t = 0.028, p = 0.489 and t = 0.016, p = 0.494, respectively). Dietary intake values can be seen in Table 1. Dietary recall rather than the Snap and Send method was used for one participant two days preceding drop out.

The demands of training (mean total distance = 3930 ± 432 m, mean accelerations = 59 ± 13, mean decelerations = 47 ± 10, mean high-speed running distance = 143 ± 94 m) and match-play (mean total distance = 6044 ± 3700 m, mean accelerations = 84 ± 35, mean decelerations = 82 ± 33, mean high-speed running = 294 ± 256 m) were captured during all sessions.

Mean glucose concentrations were 6.5 ± 1.2 mmol/L during training sessions, 7.5 ± 2.1 mmol/L during match play, and 5.4 ± 0.3 mmol/L overnight. There were no significant differences in glucose concentrations in response to match-play (p = 0.060) or training (p = 0.510), compared overnight fasted interstitial glucose concentrations. There was also no significant difference between training and match-play glucose concentrations (p = 0.788) (Figure 2).

Glucose concentrations during match-play varied between participants (Figure 2 and Table 2). During the first half of match-play the absolute change in glucose concentrations from overnight fasted glucose concentrations was between - 0.2 mmol/L and + 5.8 mmol/L and similar changes were seen during the second half of match play ( - 0.2 mmol/L and + 5.8 mmol/L).

Nicotine pouch or Snus administration (n=3) resulted in a peak increase (mean for the 3 individuals) in glucose concentration after nicotine pouch/Snus was 4.1 mmol/L. These three examples of the interstitial glucose response to nicotine pouches or Snus are demonstrated in Figure 3. These changes occurred in the absence of dietary intake and/or physical activity at the time of administration.

The primary aim of this study was to assess if the use of CGM technology was feasible and useful within professional football. This study demonstrated that CGM can be feasibly used to monitor glucose concentrations in professional football, evidenced by successful recording in 98% of the training or match play events (135 hours and 57 minutes of match play and on-pitch training sessions). Previous studies have reported limited adverse effects of CGM use, and the biometric data from CGM has been used to inform athlete fuelling and race pacing strategies 1314. Based on these findings, CGM has the potential to be used within sports to allow for bespoke fuelling strategies that could help mitigate fatigue and improve athletic performance, through immediate insights into an athlete’s physiological energy utilisation and response. The ban on CGM use in professional cycling 21 due to the technology's influence on racing seemingly supports their benefit 31. CGM also revealed some individual variability in glucose concentrations during match play and some unexpected changes in glucose concentration in response to nicotine pouch/Snus administration. In summary, these findings suggest there may be value of using CGM in professional footballers, to better understand the individual responses to nutrition and lifestyle, and identify players who may benefit from personalised support strategies.

CGM identified individual differences in blood glucose response to training and match-play. The difference between the highest and lowest glucose concentration was 5.9 mmol/L. The metabolic demands of football vary from match to match and between players 28 and this may partially contribute to the individuality observed in the present study. However, individuality has also been reported in endurance athletes 1314, where metabolic demands may be more consistent. As observed by Bowler et al., 28, day-to-day variability between athletes was more prominent than overnight variability, and secondary to diet and activity induced changes in glucose concentrations. Variability between individuals is likely to be associated with genetic differences impacting carbohydrate absorption, metabolism, uptake, and utilisation 17. Based on the variability observed in the present and previous studies, monitoring individual glucose responses using a CGM device could allow practitioners to develop personal fuelling strategies that account for individual variation in glucose availability during exercise 3031. Despite this, there is no existing data on ‘optimal’ blood glucose range for performance in football, so the value of monitoring and/or responding to glucose changes is debatable. Hiromatsu et al., (2023) concluded that it is likely to be important to maintain high glucose availability during match-play, to ensure provision of a continuous source of glucose to fuel carbohydrate metabolism, required for high-intensity efforts. However, the value of monitoring glucose concentrations via CGM for performance in football remains to be established.

CGM has improved care for people living with diabetes, with multiple studies demonstrating improvements in glycaemic control and reduced reliance on glucose-lowering medication in patients using CGM 10. Much of this success is attributed to the educational value these monitors provide patients, encouraging changes in diet and physical activity to improve glucose concentrations, leading to improved health outcomes 32. Whether continuous monitoring of glucose concentrations via CGM can influence behaviour in athletes in currently unknown. Many athletes struggle to achieve the carbohydrate intake recommendations 34, with our study finding that players did not reach the UEFA recommended carbohydrate intake of 6-8 g/kg BM 34. Interestingly we did observe a non-statistically significant tendency (p = 0.077) for carbohydrate intake to increase slightly during the CGM observation period. Although the absolute difference is small (0.2 g/kg BM), there are known limitations of food diaries to accurately capture intake 27, and as such, this finding could be interpreted as indicative of change rather than an accurate reflection of absolute intake. Further research is required in a larger sample size and over a longer time frame to better understand the potential educational value of CGM in athletes to positively influence dietary intake behaviours.

One significant consideration for practitioners surrounding the utility of CGM is that during match-play CGMs are not synchronised. This results in an mean reading being recorded every 15 minutes rather than minute by minute. There is also an approximate 2.4 ± 4.6 minutes minute delay before interstitial glucose concentrations reflect blood glucose concentrations 24, which needs to be accounted for by practitioners. However, with sufficient duration and repetition, there may be potential for CGM to generate a ‘blood glucose profile’ for individual players, which could allow for proactive nutritional approaches.

An advantage of CGM is the capturing of unexpected effects of diet or behaviour on glucose concentrations. In the present study, three of the eight participants regularly used either a tobacco free nicotine pouch or tobacco containing Snus pouch, administered between the gum and upper lip. Nicotine pouches and Snus are reported to help relaxation and reduce stress 35. A recent report found 1 in 5 professional football players within the English Football league currently use nicotine pouches or snus 35. We observed that sublabial administration of the nicotine pouch or Snus led to a mean 13% increase in mean glucose concentrations up to 55 minutes post-administration. However, peak change was variable for each administration and participant as displayed by Figure 3. Nicotine increases lipolysis and theories surrounding this response relate to it potentially being driven through the inhibition of glucose uptake 3637.

Current recommendations emphasise the importance of rapid glycogen resynthesis within the first four hours post-match, with inadequate glycogen resynthesis leading to a reduction in performance for the next 48 - 72 hours 34. As such, the novel findings from this study appear to suggest a contributing role of nicotine interfering with glucose regulation, which could influence both glucose stability and glycogen resynthesis, with the potential for negative effects on recovery. Given the widespread use of nicotine pouches and Snus in both professional and amateur football 35, further research is required to determine the effects of nicotine on blood glucose in a larger population and under controlled conditions.

Although CGM successfully recorded blood glucose on 98% of occasions, careful consideration is required on whether the performance insights gained outweigh the device cost and analysis time. However, there is value in discovering potentially unexpected changes in glucose concentrations not linked to diet or exercise, such as those seen in the current study with nicotine pouch/Snus use. Such observations may extend to pre- and post-match individualised carbohydrate recommendations and periodisation. These observations increase awareness and may enable support staff to implement preventative strategies.

CGM may increase player awareness of how diet and lifestyle practices influence their glucose concentrations, which could be used as a tool to encourage positive change. In the present study, carbohydrate intake in match day -1, was observed to be considerably lower than UEFA recommendations (observed 4.2 ± 0.9 g/kg BM, UEFA guidance 6-8 g/kg) 34, which is consistent with previous studies in football players 3839. The present study observed a modest non-significant increase in CHO intake during CGM of 0.2 g/kg BM or 20 g per day. Despite this, an argument could be made that within the applied setting this tentatively supports CGM use to positively change lifestyle behaviours, as has been previously shown in clinical populations 32.

Some issues that may impact feasibility were identified. CGM are currently targeted for clinical use and include a built-in upper limit to prevent self-diagnoses of diabetes. In the present study, one of the six participants regularly reached/exceeded the maximal reading of 11.2 mmol/L. This is a high value that would not be expected to be regularly exceeded in a professional sport context. The inability to conduct minute-by-minute analysis during training and match-play, caused by infrequent Bluetooth connection between the device and the user’s smartphone significantly limits the depth of data interpretation. It is unclear whether glucose measurements taken every 15 minutes in these dynamic environments provide valuable insights for practitioners, given the rapid fluctuations in glucose levels. More consistent, real-time monitoring may be required to fully understand the physiological changes and provide meaningful guidance for performance and health management.

Dr Ian Varley, Dr David Clayton, Ross Burbeary. Data curation – Sophie Harries

Sophie Harries

Dr Ian Varley, Dr David Clayton, Ross Burbeary Investigation – Sophie Harries

Sophie Harries, Dr Ian Varley, Dr David Clayton, Ross Burbeary Resources – Dr Ian Varley, Dr David Clayton, Dr Michael Johnson, Ross Burbeary Supervision – Dr Ian Varley, Dr David Clayton, Dr Michael Johnson, Dr Robin Thorpe Validation – Sophie Harries, Dr Ian Varley, Dr David Clayton, Dr Michael Johnson Visualization – Sophie Harries, Dr Ian Varley, Dr David Clayton, Dr Michael Johnson

Sophie Harries, Dr Ian Varley, Dr David Clayton, Dr Michael Johnson

Sophie Harries, Dr Ian Varley, Dr David Clayton, Dr Michael Johnson, Ross Burbeary, Dr Robin Thorpe.

Funding for this project was provided by Nottingham Trent University as part of a PhD studentship.

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

The data that support the findings of this study are openly available in Zenodo at https://doi.org/10.5281/zenodo.14202932

The code that support the statistical analysis of this study are openly available in Zenodo at https://doi.org/10.5281/zenodo.14202932.

Ethics approval was granted from Nottingham Trent University (1849189).