Fluid balance and electrolyte losses in collegiate male soccer players in practice and game under different environments

Abstract

The environment poses significant physiologic challenges to athletic performance. The purpose of this study was to investigate fluid balance and electrolyte losses in Chinese collegiate male soccer players during practices and games under different environments. Twelve male players were recruited across 38 practices (P) and 17 games (G) over a 9-month period, conducted in hot (H) and cool (C) environments, yielding four experimental conditions: G + H, P + H, G + C, and P + C. Fluid balance parameters comprised body mass loss (BML), urine specific gravity (USG), and sweat compositions. On-field running characteristics were collected only in games, including total distance covered (TD), number of sprints, fast run distance, accelerations, and decelerations. Pre-exercise USG indicated that > 50% of players began exercise in a hypohydrated status across four conditions. Players in G + C showed a higher proportion of ≥ 2% hypohydration, whereas players were overhydrated in G + H. Fluid balance parameters were significantly influenced by both exercise conditions and environments. Better running performance was found in G + C than G + H (p < 0.001). Both TD (r = -0.56, p < 0.001) and fast run distance (r = -0.54, p < 0.001) was inversely associated with %BML. Pre-exercise hypohydration remains a significant concern, regardless of exercise conditions or environments. TD and fast run distance may serve as practical indicators of post-game hydration status. Future guidelines aim at modifying drinking behaviors could help players maintain optimal hydration.

Introduction

Soccer is one of the most popular sports in the world, engaging millions of recreational and professional players¹. During a 90-minute game (plus stoppage time), players typically cover 9–12 km, performing a variety of movements, including running, jumping, changes in direction, accelerations, and decelerations². The exercise intensity corresponds to 80–90% of maximal heart rate; roughly 90% of the game is sustained by aerobic metabolism, while the remaining 10% relies on anaerobic energy system³. Modern soccer increasingly relies on metrics such as total distance covered and high-speed running distance, which now serve as benchmarks for distinguishing competitive levels⁴. Elite players usually cover more high-intensity running distance (~ 0.53 km) and sprint distance (0.24 km) than their less-accomplished counterparts, suggesting the growing physiological demands of the game. These higher intensities also accelerate fluid loss, making it essential for coaches and team medical staff to monitor soccer players’ hydration status and implement effective hydration strategies.

Typically, soccer players lose more fluid through sweat than they replace during games or practices, leading to ≥ 2% hypohydration^5,6. A fluid deficit of this magnitude impairs exercise performance by reducing aerobic endurance and anaerobic capacity, while also diminishing passing decision-making, and technical skills⁷. Competing in hot environments further exacerbates hypohydration, elevating core temperature, accelerating muscle glycogen depletion, and hastening fatigue. These physiological disturbances have been shown to decrease total running distance and high-intensity running distance during games, potentially altering game outcomes⁸.

Dehydration remains a persistent challenge for soccer players due to limited opportunities for fluid intake during games, which are often played in hot environmental conditions⁹. A player’s hydration status is further influenced by ambient temperature and exercise intensity, both of which should be considered when developing soccer-specific hydration strategies to mitigate hypohydration and hyponatremia risk. Therefore, the purpose of the present study was to investigate fluid balance and electrolyte losses in collegiate male soccer players during practices and games conducted in cool and hot environments. The study also sought to identify potential relationships between fluid balance and on-field running performance. We hypothesized that players would exhibit greater hypohydration risk in hot environments and during games. In addition, fluid balance would associate with on-field running performance.

Methods

Participants

Twelve collegiate male soccer players (goalkeepers excluded) from a Chinese university were recruited for this study (age: 18.8 ± 0.6 years, height: 178.1 ± 4.5 cm, and weight: 70.0 ± 4.5 kg). All participants were free from chronic disease or injuries that could impair practices and games. The study purpose and experimental procedures were instructed to the coach and players. All players were provided written informed consent before any data collection. The research protocol received ethical approval from the university’s institutional review board (#ZSRT2021062). This study was conducted in accordance with the Declaration of Helsinki.

Experimental design

The present study employed an observational field-study design where researchers maintained a non-interventional approach during practices and games throughout both pre-season and in-season. Players typically had on-field team practices 3 times per week in the afternoon (1530–1740, duration: 101 ± 20 min), consisting of warm-up, technical drills, and team scrimmage. Games were competed either in the afternoon (1530–1740) or evening (1800–1930, duration: 89 ± 20 min). A total 17 games (G) and 38 practices (P) were collected in the present study over a 9-month period. Wet-bulb global temperature (WBGT) and relative humidity (RH; AK13331) were recorded continuously and averaged for each session. According to Nassis et al.^10,11., environmental conditions were classified as follows: cool environment (C) indicating low risk for hyperthermia: WBGT < 18 °C with RH > 75%, or WBGT < 20 °C with RH = 50–75%, or WBGT < 24 °C with RH < 50%; Hot environment (H), indicating high risk for hyperthermia: WBGT > 23 °C with RH > 75%, or WBGT > 25 °C with RH = 50–75%, or WBGT > 28 °C with RH < 50%.

Fluid balance assessments

Pre- and post-exercise body mass was measured using a digital scale (TANITA HD-395, ± 0.1 kg precision) to determine body mass loss (BML), with participants instructed to void their bladder completely before each measurement. Urine samples were collected in medical grade containers and immediately analyzed for urine specific gravity (USG; ATAGO PAL-10s), with USG ≥ 1.020 indicating hypohydration¹². The USG refractometer was calibrated each day with distilled water before measurement. For sweat analysis, a standardized 16 cm² absorbent patch was affixed to the lower back. The skin was cleaned with 70% alcohol and toweling dried prior to application. Sweat patches were removed before saturation at the mid-session breaks (45–50 min) during practices and at half-time during games. Local sweat rate¹³ (LSR; mg/cm²/min) was calculated as:

After removal, each sweat patch was centrifuged to extract the sweat sample, which was then analyzed for electrolyte concentrations (sodium [Na⁺], potassium [K⁺], chloride [Cl⁻]) using an ion-selective probe electrolyte analyzer (AUDICOM, AC9900). Local sweat sodium and chloride concentration were corrected to represent the whole-body losses in accordance with Baker et al. (Na⁺: y = 0.6x + 18; Cl⁻: y = 0.62x + 19)¹⁴. Each player was provided an individually labelled water bottle and was instructed to ingest all fluids that entered the mouth. Mouth rinsing was permitted with a separate water bottle. Water bottles were weighed using a food preparation scale (TANITA KD-160BK) before and after exercise to determine total fluid intake. The team provided water only, no carbohydrate-electrolyte solution was available. Whole-body sweat rate (WBSR; L/h) was calculated from change in body mass, corrected by total fluid intake and urine output¹⁵. Tympanic temperature (T_tym) was also measured at half-time during games and at the team break during practices (Braun IRT 6030).

On-field running characteristics

Heart rate monitors (STAT Sports HR) were worn around the chest to record average heart rate (HR_ave) and maximal heart rate (HR_max), representing the physical workload during exercise. On-field running characteristics were assessed only during games with a global positioning system (GPS) tracker (STAT Sports Apex). The GPS tracker secured in a vest pocket positioned between the scapula beneath the jersey. From the GPS data, total distance covered (TD), high-speed running distance (> 14.4 km/h; HSR), sprint distance (> 21.6 km/h; SD), fast run (HSR + SD), the numbers of accelerations (ACC), and decelerations (DEC; accelerated or decelerated > 3 m/s²) were used to represent on-field running performance^16,17. Running performance variables were subsequently normalized to individual playing time (per hour).

Statistical analysis

Statistical analysis was performed using JMPro 16 (SAS Inc.), and data are expressed as mean ± standard deviation (Mean ± SD). A chi-square (X²) test was used to compare pre-exercise hydration status across the four conditions. Normality was examined with the Shapiro-Wilk test. Variables meeting the normality assumption were analyzed with a one-way ANOVA followed by Tukey’s HSD post-hoc test. Non-normally distributed variables were analyzed with the Kruskal-Wallis test, followed by non-parameter pairwise comparisons. Effect sizes for ANOVA/Kruskal-Wallis were quantified with eta square (η²): 0.01 (small), 0.06 (medium), and > 0.14 (large). The main outcomes comprised fluid balance variables (USG, %BML, water intake rate, LSR, WBSR), electrolyte losses (sweat electrolyte loss and sodium chloride loss), and physiological responses (HR_ave, HR_max, and T_tym). Game running characteristics (TD/h, HSR/h, SD/h, ACC/h, and DEC/h) were compared between G + H and G + C using independent t-tests, with Cohen’s d effect sizes (0.2 small, 0.5 medium, > 0.8 large). Pearson product moment correlations (r) examined associations between game running characteristics and fluid balance parameters. The sample size was determined using G*Power 3.1. for a one-way ANOVA analysis with four experimental groups (G + H, G + C, P + H, and P + C). Based on effect size of 0.25, an α level of 0.05, and statistical power (1-β) of 0.8, a minimum total sample size of 180 participants was required. Missing data (G + H = 10, G + C = 15, P + H = 3, P + C = 6) were excluded and final sample size detailed in Fig. 1. Statistical significance was set at p < 0.05.

Fig. 1.

Final sample size for the analysis in each condition. Sample were analyzed by soccer games (G) and practices (P) under hot (H) and cool (L) environmental conditions.

Results

Pre-exercise hydration status

Mean WBGT and RH for hot conditions were 32.5 ± 3.6 °C and 60 ± 13% during games and 33.7 ± 3.9 °C and 45 ± 18% during practices. In cool conditions, mean values were 18.2 ± 3.1 °C and 60 ± 14% during games and 16.8 ± 3.6 °C and 59 ± 13% during practices. The proportion of players arriving hypohydrated (USG ≥ 1.020) was 81.8% (G + H), 58.1% (G + C), 73.1% (P + H), and 66.1% (P + C) (X² = 8.65, p = 0.03). Pre-exercise hypohydration was higher in G + H than in G + C (X² = 7.39, p = 0.007) and P + C (X² = 4.03, p = 0.04), and higher in P + H than in G + C (X² = 4.35, p = 0.04). Regardless of exercise conditions, both pre- and post-exercise USG were significantly higher in hot than cool environments (all p < 0.001; Table 1).

Table 1.

Fluid balance and physiological variables between practices and games under different environmental conditions.

	G + H	P + H	G + C	P + C	p
Pre USG	1.024 ± 0.006 (1.022–1.026)	1.023 ± 0.007 (1.021–1.024)	1.021 ± 0.006*# (1.019–1.021)	1.020 ± 0.007*# (1.019–1.022)	< 0.001
Post USG	1.028 ± 0.003 (1.027–1.029)	1.028 ± 0.005 (1.027–1.029)	1.023 ± 0.005*# (1.021–1.024)	1.024 ± 0.006*# (1.023–1.025)	< 0.001
BML (%)	-1.07 ± 0.72 (-1.30–-0.82)	-0.76 ± 0.56*† (-0.86–-0.64)	-1.24 ± 0.79 (-1.42–-1.04)	-0.95 ± 0.49† (-1.04–-0.86)	0.001
Water intake rate (L/h)	1.18 ± 0.72 (0.94–1.40)	0.52 ± 0.30* (0.46–0.58)	0.47 ± 0.25* (0.41–0.52)	0.29 ± 0.18*#† (0.25–0.32)	< 0.001
LSR (mg/cm2/min)	3.71 ± 1.28 (3.29–4.13)	2.82 ± 1.79* (2.46–3.18)	2.51 ± 1.18* (2.24–2.78)	2.10 ± 1.21*#† (1.88–2.32)	< 0.001
WBSR (L/h)	1.74 ± 0.95 (1.41–2.06)	0.87 ± 0.36* (0.79–0.94)	1.01 ± 0.47*# (0.88–1.12)	0.65 ± 0.22*#† (0.61–0.69)	< 0.001
Sweat NaCl loss (g)	4.75 ± 2.07 (4.00–5.49)	2.58 ± 1.16* (2.32–2.83)	3.18 ± 1.33*# (2.82–3.52)	2.10 ± 0.81*#† (1.94–2.25)	< 0.001
WB-sweat Na+ (mmol/L)	51.5 ± 10.3 (48.1–54.8)	50.9 ± 10.2 (48.7–53.1)	50.2 ± 11.3 (47.6–52.9)	54.9 ± 9.4#† (53.1–56.7)	0.002
Sweat K+ (mmol/L)	3.4 ± 0.5 (3.2–3.5)	4.2 ± 1.4* (3.8–4.4)	4.1 ± 0.9* (3.9–4.4)	5.4 ± 1.9*#† (5.02–5.75)	< 0.001
WB-sweat Cl− (mmol/L)	46.8 ± 9.8 (43.6–50.0)	42.8 ± 15.1 (39.5–46.0)	46.3 ± 9.4 (44.0–48.5)	48.5 ± 8.2#† (46.9–50.1)	0.048
△ Ttym (°C)	0.92 ± 0.36 (0.81–1.03)	0.56 ± 0.44* (0.47–0.64)	0.37 ± 0.52* (0.22–0.52)	0.08 ± 0.50*#† (-0.01–0.17)	< 0.001
HRmax (bpm)	188 ± 6 (186–190)	171 ± 17* (168–175)	189 ± 10# (186–191)	183 ± 12*#† (181–186)	< 0.001
HRave (bpm)	137 ± 10 (133–140)	122 ± 16.5* (119–125)	140 ± 12.9# (136–143)	134 ± 12.0#† (132–136)	< 0.001
%HRmax	72 ± 5 (71–74)	71 ± 6 (70–72)	74 ± 5# (73–75)	73 ± 4# (72–74)	0.01

Data were presented by M ± SD (95%CI); G + H: game competed in hot environment; P + H: practiced in hot environment; G + C: game competed in cool environment; P + C: practiced in cool environment; USG-urine specific gravity; BML-body mass loss; LSR-local sweat rate; WBSR-Whole body sweat rate; NaCl-sweat sodium loss; Sweat Na⁺ and Cl⁻ presented as whole-body loss; T_tym-tympanic temperature; HR_max-maximal heart rate; HR_ave-average heart rate; ^*significant difference compared to G + H; ^#significant difference compared to P + H; ^†significant difference compared to G + C.

Exercise and environmental conditions

Overall, 88.6% of players finished with 0–2% BML. The percentages with BML > 2% were: G + H 5.3%, G + C 13.2%, P + H 1%, P + C 2.6%. Weight gain occurred in 7.9%, 1.5%, 3.8%, and 1% of sessions for G + H, G + C, P + H, and P + C, respectively.

Fluid balance and physiological variables for each condition are presented in Table 1. %BML was significantly lower in P + H (-0.76 ± 0.56%) compared to G + C (-1.24 ± 0.79%; p < 0.001, Cohen’s d = 7.4) and G + H (-1.07 ± 0.72%; p = 0.04, Cohen’s d = 3.8). P + C also showed a significantly lower %BML (-0.95 ± 0.49%) than G + C (-1.24 ± 0.79%, p = 0.01, Cohen’s d = 0.44). Water intake rate was highest in G + H and lowest in P + C compared to other conditions (p < 0.001, η² = 0.39).

G + H exhibited the greatest LSR (3.71 ± 1.28 mg/cm²/min, p < 0.001, η² = 0.11), WBSR (1.74 ± 0.95 L/h, p < 0.001, η² = 0.34), and sweat sodium chloride loss (4.75 ± 2.07 g, p < 0.001, η² = 0.31) than other conditions. Conversely, P + C showed significantly lower LSR (2.10 ± 1.21 mg/cm²/min) than P + H (2.82 ± 1.79 mg/cm²/min, p = 0.004, Cohen’s d = 0.47), and also showed significantly lower WBSR (0.65 ± 0.22 L/h) than P + H (0.87 ± 0.36 L/h, p < 0.001, Cohen’s d = 0.74) and G + C (1.01 ± 0.47 L/h, p < 0.001, Cohen’s d = 0.98). G + C also produced greater sodium chloride loss (3.18 ± 1.33 g) compared to P + H (2.58 ± 1.16 g, p = 0.007, Cohen’s d = 0.48) and P + C (2.10 ± 0.81 g, p < 0.001, Cohen’s d = 0.98). Sweat electrolyte concentrations revealed that P + C had significantly greater whole-body sweat Na⁺ (54.9 ± 9.4 mmol/L) and Cl⁻ concentration (48.5 ± 8.2 mmol/L) compared to P + H (Na⁺: 50.9 ± 10.2 mmol/L, p = 0.003, Cohen’s d = 0.41; Cl⁻: 42.8 ± 15.1 mmol/L, p = 0.01, Cohen’s d = 0.34) and G + C (Na⁺: 50.2 ± 11.3 mmol/L, p = 0.001, Cohen’s d = 0.45; Cl⁻: 46.3 ± 9.4 mmol/L, p = 0.03, Cohen’s d = 0.25). G + H showed the lowest sweat K⁺ concentration (3.4 ± 0.5 mmol/L) compared to other conditions (p < 0.001, η² = 0.20), whereas P + C showed the highest K⁺ concentration (5.4 ± 1.9 mmol/L) compared to other conditions (p < 0.001, η² = 0.20).

The rise in T_tym was significantly greater in G + H (0.92 ± 0.36 °C) compared to other conditions (all p < 0.001, η² = 0.29). In addition, players in P + C had significantly lower T_tym elevation (0.08 ± 0.50 °C) compared to P + H (0.56 ± 0.44 °C, p < 0.001, Cohen’s d = 1.02) and G + C (0.37 ± 0.52 °C, p = 0.004, Cohen’s d = 0.57). The rise in T_tym was significantly associated with water intake rate (r = 0.40, p = 0.001). Players in P + H exhibited the lowest HR_max (171 ± 17 bpm, p < 0.001, η² = 0.22) and HR_ave (122 ± 16.5 bpm, p < 0.001, η² = 0.21) than other conditions, but did not differ between cool and hot games.

On-field running characteristics

On-field running characteristics under hot and cool conditions are presented in the Table 2. Players competed in cool conditions had significantly greater TD/h, HSR/h, SD/h, fast run%, numbers of ACC/h, and numbers of DEC/h than in hot conditions (Fig. 2). Furthermore, %BML was significantly associated with TD (r = -0.56, p < 0.001) and fast run (r = -0.54, p < 0.001, Fig. 3).

Table 2.

On-field running characteristics during soccer games.

	G + H	G + L	p	Cohen’s d
TD (m/h)	4559 ± 1453	5863 ± 772	< 0.001	1.12
HSR (m/h)	158 ± 140	283 ± 107	< 0.001	1.00
SD (m/h)	111 ± 105	182 ± 94	0.003	0.71
Fast run%	5.5 ± 3.9	7.9 ± 2.8	0.003	0.71
ACC (#/h)	29 ± 12	39 ± 11	< 0.001	0.87
DEC (#/h)	27 ± 15	44 ± 12	< 0.001	1.25

TD-total distance covered; HSR-high-speed running distance (> 14.4 km/h); SD-sprint distance (> 21.6 km/h); Fast run%-the proportion of HSR and SD to the TD. ACC: number of accelerations; DEC: number of decelerations.

Fig. 2.

Sprint distance (A), HSR (high-speed running distance, B), and total distance covered (C) during games in hot and cool environments. *Significant difference in cool compared to hot environments. G + H: game competed in hot environment; P + H: practiced in hot environment; G + C: game competed in cool environment; P + C: practiced in cool environment.

Fig. 3.

Correlations between %BML with total distance (A) and fast run (B) in games. Total distance was significantly associated with %BML (r = -0.56, p < 0.001); fast run distance indicates the sum of sprint distance and high-speed running distance and was significantly associated with %BML (r = -0.54, p < 0.001).

Discussion

This study determined fluid balance, electrolyte losses, and physiological responses of collegiate male soccer players during practices and games under both hot and cool environments. The present study also investigated the relationships between hydration status and game running characteristics. The main findings revealed that: (1) the majority of players began exercise in a hypohydration state, with rates highest before games in the heat (~ 82%); (2) exercise conditions (practice vs. game) and environments (hot vs. cool) significantly influenced fluid balance and physiological variables, with games in hot environments eliciting the greater thermal load and the poorer performance; and (3) during games, both total distance covered and fast run distance were inversely associated with %BML.

Pre-exercise hydration status

Pre-exercise hydration status is important to maintain optimal exercise performance yet frequently neglected by players. A recent meta-analysis revealed that pre-exercise hypohydration significantly compromises aerobic performance (-2.4%), peak oxygen consumption (-2.4%), and oxygen uptake at lactate threshold (-4.4%)¹⁸. This study observed 58.1–81.8% of players arriving hypohydrated, aligning with previous reports in soccer populations^19,20. Although, the American College of Sports Medicine (ACSM) emphasizes pre-exercise hydration and recommends consuming 5–7 ml/kg and 3–5 ml/kg of fluid 4 h and 2 h prior to exercise, respectively²¹, the persistent prevalence of hypohydration among soccer players suggests inadequate education. Li et al. identified a health-psychology framework that integrates self-determination theory and the theory of planned behavior to examine the pre-exercise fluid intake behavior of Chinese collegiate players. They found that such behavior is associated with both motivational and social-cognitive factors within the model, suggesting interventions should target all of the relevant factors to enhance player’s intentions and actual pre-exercise fluid intake behavior²². The present soccer team does not have athletic trainers or medical staff, whereas head coach has more attention on training and skills development. In addition, most players in this study were freshmen who had never received any nutrition or hydration education. Thus, those players may not be aware of how pre-exercise hydration influences subsequent performance. An effective hydration education model should be integrated into training programs and appropriate hydration knowledge is necessary to deliver to the players, such as practical self-assessment tools (i.e., morning monitoring body mass or urine color) and highlighting carbohydrate-electrolyte solutions to improve fluid retention.

Exercise and environmental conditions

Fluid balance and physiological responses varied significantly depending on exercise conditions and environments. While most players maintained 0–2% BML during both practices and games. However, a higher proportion exceeded 2% BML during games in cool environments. This finding contrast with previous reports suggesting > 2% BML occurs more frequently in hot conditions^7,23. The observed discrepancy may be explained by players wearing additional clothing in cool conditions, which increased sweat loss during high intensity efforts^24,25. This phenomenon similarly documented in ice hockey players wearing protective equipment at ~ 10 ℃ ambient temperature¹⁵. In addition, fluid replacement patterns showed environmental dependence of the present study. Players replaced only 42.3% of sweat loss in G + C, while intake rates were significantly lower in cool environments (0.47 L/h in G + C and 0.29 L/h in P + C). This aligns with established evidence that fluid intake typically increases above 25℃⁹, and that replacing < 50% of fluid losses elevates hypohydration risk (≥ 2% BML)²⁶. The reduced intake in cool conditions likely reflects both diminished thermal stress and attenuated thirst perception. Notably, three players consumed no fluids during G + C and another player who drank nothing during three separate practice sessions. When we asked them why they do not drink, these players replied similarly “not feeling thirst and drinking needs.” Thus, deliberate fluid-replacement strategies remain essential even when competing in cool conditions.

Notably, several players consumed excessive water and finished exercise with body-mass gain, particularly during games in hot environments. This over-consumption likely stems from the combination effects of pre-exercise hypohydration and thermal stress. Pre-exercise hypohydration may induce thirst perception while blunting sweat sensitivity during subsequent exercise²⁷. However, ingesting large volumes of plain water remains problematic because it elevates exercise-associated hyponatremia risk among high-sodium sweaters²⁸. This is a particular concerns given the team only provided water without electrolyte replacement. ACSM guidelines specifically recommend 20–30 mmol/L sodium-containing carbohydrate-electrolyte beverages for exercise sessions exceeding one hour²¹, underscoring the need for provision of such solutions in future exercise. On the other hand, water intake rate varied significantly across conditions, with player consuming 0.71 L/h more during G + H than G + C despite comparable exercise intensities. This disparity appears thermally mediated, as G + H experienced a 0.92 ℃ T_tym elevation versus 0.37 ℃ in G + C. Our data also revealed significant correlation between water intake rate and rises in T_tym, but not with exercise intensity. This results contrast with Rollo et al.’s findings in elite European players who maintained consistent intake across conditions except during low-intensity exercise in cool conditions²⁹. This may be explained by physiological adaptations in elite players (superior aerobic capacity and thermal tolerance) compared to our colligate cohort. Both hypohydration and hyperhydration ultimately derive from inappropriate fluid intake management. Our players mainly rely on thirst to drink, as they experience lower thirst, they do not realize the importance of hydration maintaining exercise performance. Large volume water ingested as thermal stress elevated.

This study showed that an average predicted whole-body sweat sodium concentration of 52.2 ± 10.4 mmol/L (range 31.2–81.1 mmol/L) across all conditions. These values are higher than those previous reported in Suarez-Ortegón et al., (43 ± 15 mmol/L) and Shirreffs et al., (30.2 ± 18.8 mmol/L)^30,31. Although methodological differences in sweat collection sites (forearm vs. lower back) may contribute to this discrepancy, sweat-sodium concentration is also influenced by diet, exercise intensity, environmental stress, and heat-acclimation status, making cross-study comparisons difficult^32,33. However, a large inter-individual variability (31.6 ~ 81.2mmol/L) was observed in the present study, consistent with previous reported in soccer and other sports (12.6 ~ 104.8mmol/L)^32,34. This suggests that general hydration guidelines cannot accommodate such individual differences to meet each individual’s needs. A recent study demonstrated that personalized hydration strategies incorporating sodium content and fluid volume, which optimized hydration status and enhanced intermittent exercise performance in the heat³⁵. We advocate implementing the current protocol throughout the season to develop personalized sodium-replacement strategies that simultaneously optimize performance while mitigating risks of both dehydration and hyponatremia.

On-field running performance

The present finding regarding on-field running performance in collegiate soccer players may not be directly comparable to the previous data from professional athletes. Our analysis showed that player’s running distance at different speed, number of accelerations and decelerations during games were significant lower under hot compared to cool environments. Previous research demonstrated that a 7% reduction in total distance covered and 26% decrease in high-intensity running distance in hot conditions compared to temperate environments³⁶. In contemporary soccer, players are required to cover greater total distance and high intensity running distance for winning³⁷. Importantly, we identified a moderate correlation between total distance and fast run distance with %BML during games. Our linear regression model predicted that players covering an average of 5.8 km total distance and 0.34 km fast run distance would experience − 1.0% and − 1.1%BML, respectively. This model also helps to explain observed hyperhydration cases. Those players who gained body mass ran 4.3 km in total distance and 0.36 km fast run should have theoretically lost 0.7–1.1%BML. Based on their pre-exercise body mass (68.6 kg), appropriate fluid intake should have ranged 480–755 g, yet actual consumption reached 1452 g, clearly indicating over consumption. These findings support Sekiguchi et al. suggesting total distance covered serves as a reliable predictor of fluid losses³⁸. While coaches and sport scientists recognize running metrics as key performance indicators, the connection to hydration status remains underappreciated. Our results suggest that when direct hydration assessment is impractical during soccer games, running performance data can provide valuable estimates of fluid balance. This approach may help prevent both dehydration and the potentially inappropriate overconsumption observed in our study.

Strength and limitations

Previous investigations of soccer hydration have primarily examined only hot or cool environments with limited observations (typically one or few sessions), restricting comprehensive comparisons of multiple factors affecting fluid balance in real-world conditions. The present study investigated in a longitudinal observational design that tracked the same cohort across multiple practices and games in both hot and cool environments. However, several limitations should be addressed in the present study, with some drop cases in each condition and not all the players participated in the game according to the coach decisions. In addition, our participants were Chinese collegiate player, generalizability to other populations (e.g., professionals, different ethnicities) may be limited. Lastly, urine sample temperature was not standardized for USG measurements, which could potentially affect readings³⁹.

Practical applications

Coaching staff should systematically identify and address barriers to pre-exercise hydration while implementing practical education intervention, particularly for developing players. Regular on-field hydration monitoring during training sessions is essential to establish individual fluid balance patterns and electrolyte loss profiles, especially due to temperature variation in a year-long season. Team physicians and coaches must identify those players who have inadequate or over fluid consumption during practices and games. Total distance covered and fast run during games may be a feasible point for educating players on monitoring their personal hydration status.

Conclusions

Pre-exercise hydration status persists as a concern for players and performance, regardless of exercise or environmental conditions. In addition, soccer players exhibited different fluid balance in practices and games across hot and cool environments. Thermal stress appears to be a primary driver of drinking behaviors, with substantially higher fluid consumption observed in hot conditions compared to markedly reduced intake in cool environments. Total distance covered and fast run distance can be seen as practical predictors of hydration status when on-field hydration assessments are unavailable. These results highlight the need for environment-specific hydration strategies while demonstrating the potential utility of performance data for monitoring fluid balance during games.

Author contributions

H.W. contributed to the conception and design of this study. P.M., G.Z., and L.H. contributed to the data collection. H.W. and K.E. contributed to the data interpretation. P.M. wrote the original manuscript. H.W. and K.E. critically revised and edited the manuscript. All the authors have read and agreed to the published version of the manuscript.

Funding

The study was founded by the University Doctoral Research Starting Funds (#YS304320150), Zhejiang Province Master Student Research Funds (#KYZ04Y22296), Department of Education of Zhejiang Province Funds (#Y202455456), and Sports Science Innovation Project of General Administration of Sport of China (#23KJCX050).

Data availability

The data presented in this study are available on request from the corresponding author.

Declarations

Competing interests

The authors declare no competing interests.