Application of Continuous Glucose Monitoring for Assessment of Individual Carbohydrate Requirement during Ultramarathon Race

Background: The current study intended to evaluate the feasibility of the application of continuous glucose monitoring to guarantee optimal intake of carbohydrate to maintain blood glucose levels during a 160-km ultramarathon race. Methods: Seven ultramarathon runners (four male and three female) took part in the study. The glucose pro le was monitored continuously throughout the race, which was divided into 11 segments by timing gates. Running speed in each segment was standardized to the average of the top ve nishers for each gender. Food and drink intake during the race were recorded and carbohydrate and energy intake were calculated. Results: Observed glucose levels ranged between 61.9–252.0 mg/dL. Average glucose concentration di ered from the start to the end of the race (104 15.0 to 164 30.5 SD mg/dL). The total amount of carbohydrate intake during the race ranged from 0.27 to 1.14 g/kg/h. Glucose concentration positively correlated with running speeds in segments (P<0.005). Energy and carbohydrate intake positively correlated with overall running speed (P<0.01). Conclusion: The present study demonstrates that continuous glucose monitoring could be practical to guarantee optimal carbohydrate intake for each ultramarathon runner. Keywords:sports nutrition; continuous glucose monitoring; carbohydrate; trail running; Freestyle Libre 1. Introduction For the rst time in human history, in 2019, Eliud Kipchoge ran the marathon distance in under two hours. Recent advances in the area of sports science signi cantly contributed to his success. In terms of exercise nutrition, it has been

optimal carbohydrate intake for each ultramarathon runner. Keywords:sports nutrition; continuous glucose monitoring; carbohydrate; trail running; Freestyle Libre 1. Introduction For the rst time in human history, in 2019, Eliud Kipchoge ran the marathon distance in under two hours. Recent advances in the area of sports science signi cantly contributed to his success. In terms of exercise nutrition, it has been recommended to consume 90 g/h of carbohydrates for endurance exercise [1,2]. This amount has been suggested based on the maximum oxidation of carbohydrate as an energy substrate [3,4] and it is noted that the rate-limiting step to oxidizing this amount of carbohydrate is the gastrointestinal absorption process [1]. A longer distance marathon is known as an ultramarathon, and the popularity of these events has increased in recent years [5]. The total energy expenditure of a 160 km ultramarathon reaches about 13,000 kcal [6]. Thus, nutritional strategies have to be considered for ultramarathon runners wanting to improve their race results, but also for those focusing primarily on nishing the event. GI distress, which is frequently experienced by runners during all types of endurance exercise, makes the current carbohydrate intake recommendation di cult to achieve [7–10]. Several observation studies have shown that carbohydrate intake during ultramarathon races is lower than the current Nutrients2020,12, 1121; doi:10.3390 /nu12041121 /journal/nutrients

Nutrients2020,12, 1121 2 of 11 recommendation for carbohydrate intake. In addition to these statements and recommendations, the optimal nutritional strategies for ultramarathons have been proposed based on a baseline metabolic model [11]. It has been reported that only one study [12] achieved the carbohydrate amount suggested in the current recommendation, while others achieved less than the 60 g/h lower level of the recommendation. The lowest observed average was 31 g/h in slower runners [13]. A recently published position statement of the International Society of Sports Nutrition recommended the consumption of 150–400 kcal/h (carbohydrate, 30–50 g/h) [9]. Recent practical recommendations for ultramarathon events o ered advice to consume tolerable carbohydrate intake quantities during exercise, which corresponded to 0.8–1.0 g/kg/h of carbohydrate [14]. These values were provided by comparing the race diet between fast and slow runners [13] or by comparing the carbohydrate intake of nishers and non- nishers [12]. Optimal nutrition results in a decreased risk of energy depletion, better performance [10], the prevention of acute cognitive decline, and improved athlete safety on ultramarathon courses with technical terrain or those requiring navigation [9]. However, it may prove di cult for the runner to execute the precise nutrition plan [11] and the carbohydrate requirement for ultramarathon racing varies greatly depending on the individual [9]. The aim of this study was to evaluate the feasibility of continuous glucose monitoring to improve the carbohydrate intake of ultrarunners using a continuous glucose monitoring system [15,16]. 2. Materials and Methods 2.1. Study Design This observational study was designed to determine the minimum carbohydrate requirement to maintain blood glucose level and race speed during ultramarathons. All procedures were approved by the Ryukoku University Human Research Ethics Review Board (No. 2016-08-02). All research procedures complied with the code of ethics of the World Medical Association (Declaration of Helsinki). Written informed consent was obtained from all the participants before the commencement of the study. 2.2. Study Population Seven runners (4 male and 3 female) without injuries volunteered to participate in the study. All the runners had completed 2 to 3 races certi ed by the International Trail Running Association

of ethics of the World Medical Association (Declaration of Helsinki). Written informed consent was obtained from all the participants before the commencement of the study. 2.2. Study Population Seven runners (4 male and 3 female) without injuries volunteered to participate in the study. All the runners had completed 2 to 3 races certi ed by the International Trail Running Association and the sum of nisher's points exceeded 12 in the last 3 years, demonstrating their experience in running Ultramarathons. Participant characteristics are presented in Table. Table 1.Clinical characteristics of male and female subjects. Male Female P Age (year) 41.5 6.2 42.6 1.2 0.627 Height (cm) 172.9 2.7 158.0 6.5 0.019 Weight (kg) 66.0 9.3 47.9 3.8 0.036 BMI (kg/m 2 ) 22.2 2.8 18.9 0.7 0.116 Lean body mass (kg) 56.3 5.9 40.7 4.3 0.012 Fat mass (kg) 22.2 2.8 18.9 0.7 0.116 Values are means SD (male,n=4; female,n=3). 2.3. Race Course The present study was conducted during the 2019 Ultra trail Mt. Fuji (https: //www.ultratrailmtfuji. com/), held during the last week of April, around Mt. Fuji in Japan (ambient temperature range: 2.3–19.9 C). The distance of the course covered 165 km and the total elevation was 7942 m. The course included trails, rocks, paths, grasslands, and pavements. The course was divided into 11 segments by 10 timing gates where each runner's passing time was recorded electronically. Distances between each

Nutrients2020,12, 1121 3 of 11 timing gate were 15 5.4 SD km and varied from 7 to 28 km. Running time and speed between each timing gate were obtained from the o cial race web site. Running time between each timing gate was 1:58 0:48 and 2:18 0:52 h:m for the top 5 male and female nishers, respectively. All the runners had to run with backpacks to carry necessities, including food, and they could replenish food and uid at each timing gate. 2.4. Running Speed Data Collection and Standardization Running speed between each timing gate and overall running speed were obtained from the o cial race web site. The standard running speed of male and female participants (designated as 100%) for each segment were calculated by averaging the top ve male and female nishers, respectively. The running speed of subjects in each segment was standardized using the following formula. The standardized running speed exceeds 100% only when running at a pace comparable to the top 1 and 2 places in each gender: %Running speed=(The subject's running speed)/((Average of top 5 nishers' running speed in each gender)) 100, (1) 2.5. Glucose Data Collection and Standardization Blood glucose pro le was monitored by a minimally invasive method known as ash glucose monitoring (FGM). Its details have been reported elsewhere [15,17,18]. Brie y, the FGM system (FreeStyle Libre; Abbott Diabetes Care, Alameda, CA) mechanically reads and continuously measures glucose concentration in the interstitial uid collected from cells immediately below the skin and produces the corresponding ambulatory glucose pro le. Subjects were asked to attach the device more than 1 day before the race. The FGM sensor was applied at the back of the upper arm and glucose concentrations were obtained every 15 min [17]. The glucose concentration of each runner during the race was standardized by subtracting the resting fasting glucose concentration of the runner and was expressed as an increase from resting fasting glucose level (Dglucose). The average, highest, lowest, and the di erence between the highest and lowest levels ofDglucose in each segment were used as representative values in each segment

[17]. The glucose concentration of each runner during the race was standardized by subtracting the resting fasting glucose concentration of the runner and was expressed as an increase from resting fasting glucose level (Dglucose). The average, highest, lowest, and the di erence between the highest and lowest levels ofDglucose in each segment were used as representative values in each segment (Figure).Nutrients 2020, 12, x FOR PEER REVIEW 3 of 12 2.3. Race Course The present study was conducted duri ng the 2019 Ultra trail Mt. Fuji (https://www.ultratrailmtfuji.com/), held during the last week of April, around Mt. Fuji in Japan (ambient temperature range: 2.3–19.9˚C). The distance of the course covered 165 km and the total elevation was 7942 m. The course included trails, rocks, paths, grasslands, and pavements. The course was divided into 11 segments by 10 timing gates where each runner's passing time was recorded electronically. Distances between each timing gate were 15 ± 5.4 SD km and varied from 7 to 28 km. Running time and speed between each timing gate were obtained from the official race web site. Running time between each timing gate was 1:58 ± 0:48 and 2:18 ± 0:52 h:m for the top 5 male and female finishers, respectively. All the runners had to run with backpacks to carry necessities, including food, and they could replenish food and fluid at each timing gate. 2.4. Running Speed Data Collection and Standardization Running speed between each timing gate and overall running speed were obtained from the official race web site. The standard running speed of male and female participants (designated as 100%) for each segment were calculated by averaging the top five male and female finishers, respectively. The running speed of subjects in each segment was standardized using the following formula. The standardized running speed exceeds 100% only when running at a pace comparable to the top 1 and 2 places in each gender: %Running speed = (The subject's running speed) / ((Average of top 5 finishers' running speed in each gender)) × 100, (1) 2.5. Glucose Data Collection and Standardization Blood glucose profile was monitored by

using the following formula. The standardized running speed exceeds 100% only when running at a pace comparable to the top 1 and 2 places in each gender: %Running speed = (The subject's running speed) / ((Average of top 5 finishers' running speed in each gender)) × 100, (1) 2.5. Glucose Data Collection and Standardization Blood glucose profile was monitored by a minimally invasive method known as flash glucose monitoring (FGM). Its details have been reported elsewhere [15,17,18]. Briefly, the FGM system (FreeStyle Libre; Abbott Diabetes Care, Alameda, CA) mechanically reads and continuously measures glucose concentration in the interstitial fluid collected from cells immediately below the skin and produces the corresponding ambulatory glucose profile. Subjects were asked to attach the device more than 1 day before the race. The FGM sensor was applied at the back of the upper arm and glucose concentrations were obtained every 15 minutes [17]. The glucose concentration of each runner during the race was standardized by subtracting the resting fasting glucose concentration of the runner and was expressed as an increase from resting fasting glucose level (Δglucose). The average, highest, lowest, and the difference between the highest and lowest levels of Δglucose in each segment were used as representative values in each segment (Figure 1). Figure 1. Schematic presentation of the standardization of glucose levels during the race. The overall race course was divided into 11 segments (arrows) by 10 timing gates. The altitude profile of the race Figure 1. Schematic presentation of the standardization of glucose levels during the race. The overall race course was divided into 11 segments (arrows) by 10 timing gates. The altitude profile of the race course (filled area) and the change of glucose level (solid line) of the first 12 h of the race is shown as a representative result.DGlucose level was obtained by subtracting the resting fasting glucose concentration of each runner (dashed line). *, highest value ofDglucose in each segment;y, lowest value ofDglucose in each segment; dotted line, average value ofDglucose in each segment. Running speed (%) was calculated by dividing each runner's running speed by the average

h of the race is shown as a representative result.DGlucose level was obtained by subtracting the resting fasting glucose concentration of each runner (dashed line). *, highest value ofDglucose in each segment;y, lowest value ofDglucose in each segment; dotted line, average value ofDglucose in each segment. Running speed (%) was calculated by dividing each runner's running speed by the average running speed of top 5 finishers.

Nutrients2020,12, 1121 4 of 11 2.6. Diet Supply Data Collection Runners were asked to record their entire food and drink intake throughout the race. They reported the timing and volume of consumed food products and uids based on pictures taken throughout the race. Food products and uids consumed more than 60 min before the race start were not included in the calculation of nutritional intake. The energy and carbohydrate intake during the race were calculated based on the nutrition information provided by manufacturers. If data was not available, intakes were calculated based on the standard tables of food composition in Japan 2015 - (7th revised edition) [19]. The energy and carbohydrate intake were expressed relative to kg of pre-race body weight, per hour of running time. All foods were categorized with reference to previous research [20] as: sports drinks (isotonic and hypertonic formulas), gels, cola, other uids (all other drinks consumed), sweets, fruits, bars, noodles, bread, rice products and other solids (all other products consumed). 2.7. Statistics The data reported in the text, tables, and gures are presented as means and standard deviations, unless otherwise speci ed. Data were processed and analyzed in GraphPad Prism for Mac (version 8.3.1, GraphPad Inc., San Diego, CA, USA). Pearson's correlation coe cients were used to investigate the associations between running speed, glucose level, and carbohydrate intake. One-way ANOVA followed by Tukey's post-hoc test were used to compare the di erences between each runner's blood glucose level. Results were considered signi cant whenP<0.05. 3. Results 3.1. General Results The running speed of the participants ranged from 3.90 to 7.22 km/h with a standardized running speed ranging from 49.0% to 90.1%. 3.2. Relationship between Glucose Level and Running Speed All participants were within the expected normoglycemic range during exercise (72–252 mg/dL) with the exception of one participant who exhibited a lowest value of 61.9 mg/dL as shown in Table Carbohydrate mainly supplied total energy intake during the race (77.6 8.58SD% of total energy intake). Table 2.The total energy and nutrient intake, and glucose concentration during the ultramarathon. Subject FS 1 2 3 MS 4 5

normoglycemic range during exercise (72–252 mg/dL) with the exception of one participant who exhibited a lowest value of 61.9 mg/dL as shown in Table Carbohydrate mainly supplied total energy intake during the race (77.6 8.58SD% of total energy intake). Table 2.The total energy and nutrient intake, and glucose concentration during the ultramarathon. Subject FS 1 2 3 MS 4 5 6 7 Sex F F F M M M M M Running speed (%) 100 89.5 87.9 72.9 100 90.1 70.0 62.0 49.0 (min/km)6.37 5.70 5.60 4.64 8.01 7.22 5.60 4.96 3.90 Energy intake (kcal/kg/h) - 5.40 4.79 1.91 - 4.37 3.03 1.41 1.46 Carbohydrate intake (g/kg/h) - 1.14 1.04 0.34 - 0.85 0.64 0.27 0.28 Protein intake (g/kg/h) - 0.132 0.061 0.042 - 0.143 0.051 0.021 0.021 Fat intake (g/kg/h) - 0.037 0.040 0.046 - 0.040 0.036 0.030 0.029 Glucose (mg/dL) During race Average - 131 137 104 - 145 134 121 164 SD - 11.9 30.2 15.0 - 20.4 20.2 22.9 30.5 Highest - 173 224 151 - 193 198 189 240 Lowest - 105 79 62 - 100 94 83 103 Resting fasting - 53 58 40 - 57 68 83 98 FS and MS, female and male standard running speed, which correspond to the average of the top 5 nishers in each sex. F, female; M, male. Each runner consumed carbohydrates from liquids, gels, fruits, sweets or solids as shown in Table Six of 7 runners consumed more than 55% of their carbohydrates from liquids and gels (55.3% to 74.8%)

Nutrients2020,12, 1121 5 of 11 except for one runner (28.4%, subject 3). Carbohydrate intake from solids ranged from 21.1% to 42.8% in the six runners and 63.8% in the other runner, who showed the highest fat intake among 7 runners (subject 3). Table 3.Carbohydrates consumed per product type (g/kg/h). Subject 1 2 3 4 5 6 7 % of Total Liquids and gels 0.85 0.71 0.10 0.57 0.35 0.16 0.16 58. 8 15.1 Sports drink 0.34 0.05 0.04 0.01 0.00 0.05 0.04 11.8 10.8 Cola 0.00 0.08 0.00 0.05 0.03 0.06 0.02 6.9 7.5 Gel 0.51 0.57 0.04 0.51 0.32 0.05 0.06 37.2 19.6 Other liquid 0.00 0.01 0.01 0.01 0.00 0.01 0.03 3.0 4.3 Fruits and sweets 0.05 0.06 0.03 0.01 0.06 0.00 0.00 4.2 3.7 Fruit 0.04 0.04 0.02 0.01 0.06 0.00 0.00 3.8 3.5 Sweet 0.01 0.02 0.00 0.00 0.00 0.00 0.00 0.4 0.7 Solids 0.24 0.27 0.21 0.27 0.23 0.10 0.12 37.0 13.9 Bar 0.05 0.00 0.00 0.05 0.00 0.00 0.00 1.6 2.7 Noodle 0.02 0.02 0.08 0.00 0.00 0.01 0.00 4.2 8.1 Bread 0.00 0.05 0.05 0.01 0.04 0.04 0.03 7.9 6.5 Rice product 0.00 0.19 0.09 0.18 0.15 0.04 0.08 18.9 9.7 Other solid 0.17 0.01 0.00 0.04 0.03 0.01 0.00 4.4 4.9 Total 1.14 1.04 0.34 0.85 0.64 0.27 0.28 Subject numbers are identical to Table. The subtotal of each category is shown in bold. The average, highest, lowest, and the difference between the highest and lowest levels ofDglucose in 11 segments were subjected to correlation analysis between running speed and blood glucose level. Figure (r 2 =0.2397,P=0.0028; r 2 =0.1397,P=0.0501 for male and female, respectively) and average (r 2 =0.1650, P=0.0155; r 2 =0.0531,P=0.2381 for male and female, respectively) levels ofDglucose had a significant positive correlation with running speed, but not for the highest levels ofDglucose (r 2 =0.0005,P=0.8952; r 2 =0.0125,P=0.5704 for male and female, respectively) in male runners. Similar but not significant tendencies were observed in female runners. Interestingly, a significant inverse correlation (r 2 =0.1198, P=0.0417; r 2 =0.0107,P=0.6011 for male and female, respectively) was observed between running speed

a significant positive correlation with running speed, but not for the highest levels ofDglucose (r 2 =0.0005,P=0.8952; r 2 =0.0125,P=0.5704 for male and female, respectively) in male runners. Similar but not significant tendencies were observed in female runners. Interestingly, a significant inverse correlation (r 2 =0.1198, P=0.0417; r 2 =0.0107,P=0.6011 for male and female, respectively) was observed between running speed and the difference between highest and lowest (D) in male runners.Nutrients 2020, 12, x FOR PEER REVIEW 6 of 12 Figure 2. Scatter plots showing relationships between glucose level and running speed. The lowest (A), average (B), highest (C), and difference between highest and lowest (D) value of Δglucose levels were calculated as described in Figure 1. Each plot indicates one segment. 3.3. Relationship between Energy and Carbohydrate Intake and Running Speed Energy intake exhibited a significant positive correlation with running speed (r 2 = 0.8142, P = 0.0054). Energy intake ranged from 1.41 to 5.40 kcal/kg/h, which is the equivalent of 86.2 to 226.7 kcal/h. A significant correlation was also found between carbohydrate intake and running speed (r 2 = 0.7955, P = 0.0070). Carbohydrate intake ranged from 0.27 to 1.14 g/h/kg (1.1 to 4.6 kcal/h/kg), which is the equivalent of 16.3 to 52.9 g/h. The energy intake from carbohydrates contributed 63% to 87% of the total energy consumed during the race. No significant correlations were observed between running speed and energy intake from protein and fat (Figure 3). 0 50 100 150 20 40 60 80 100 120 Lowest Δglucose (mg/dL) Running speed (%) Male, Female r 2 = 0.2397, 0.1397 P = 0.0028, 0.0501 Male Female 0 50 100 150 20 40 60 80 100 120 Highest Δglucose (mg/dL) Running speed (%) Male Female Male, Female r 2 = 0.0005, 0.0125 P = 0.8952, 0.5704 0 50 100 150 20 40 60 80 100 120 Average Δglucose (mg/dL) Running speed (%) Male Female Male, Female r 2 = 0.1650, 0.0531 P = 0.0155, 0.2381 0 50 100 150 20 40 60 80 100 120 Highest - lowest Δglucose (mg/dL) Running speed (%) Male Female Male, Female r

This research assesses carbohydrate requirements for ultramarathon runners using continuous glucose monitoring.