Whole-Body Metabolism, Carbohydrate Utilization, and Caloric Energy Balance After Sport Concussion: A Pilot Study

Abstract

Background:

Sport concussion (SC) causes an energy crisis in the brain by increasing energy demand, decreasing energy supply, and altering metabolic resources. Whole-body resting metabolic rate (RMR) is elevated after more severe brain injuries, but RMR changes are unknown after SC. The purpose of this study was to longitudinally examine energy-related changes in collegiate athletes after SC.

Hypothesis:

RMR and energy consumption will increase acutely after SC and will return to control levels with recovery.

Study Design:

Case-control study.

Level of Evidence:

Level 4.

Methods:

A total of 20 collegiate athletes with SC (mean age, 19.3 ± 1.08 years; mean height, 1.77 ± 0.11 m; mean weight, 79.6 ± 23.37 kg; 55% female) were compared with 20 matched controls (mean age, 20.8 ± 2.17 years; mean height, 1.77 ± 0.10 m; mean weight, 81.9 ± 23.45 kg; 55% female). RMR, percentage carbohydrate use (%CHO), and energy balance (EBal; ratio between caloric consumption and expenditure) were assessed 3 times: T₁, ≤72 hours after SC; T₂, 7 days after T₁; and T_F, after symptom resolution. A 2 × 2 × 3 (group × sex × time) multivariate analysis of variance assessed RMR, %CHO, and EBal. Changes in RMR, %CHO, and EBal (T₁ to T_F) were correlated with days to symptom-free and days to return to play in the concussed group.

Results:

Women reported being symptom-free (median, 6 days; range, 3-10 days) sooner than men (median, 11 days; range, 7-16 days). RMR and %CHO did not differ across time between groups or for group × sex interaction. SC participants had higher EBal than controls at T₁ (P = 0.016) and T₂ (P = 0.010). In men with SC, increasing %CHO over time correlated with days to symptom-free (r = 0.735 and P = 0.038, respectively) and days to return to play (r = 0.829 and P = 0.021, respectively).

Conclusion:

Participants with SC were in energy surplus acutely after injury. Although women recovered more quickly than men, men had carbohydrate metabolism changes that correlated with recovery time.

Clinical Relevance:

This pilot study shows that male and female student-athletes may have differing physiologic responses to SC and that there may be a role for dietary intervention to improve clinical outcomes after SC.

Sport concussion (SC) is often described as a “neurometabolic cascade” that occurs subsequent to trauma.¹² This neurometabolic cascade results in an intracranial metabolic disturbance characterized by increased energy demand, decreased total energy supply, and a subsequent shift in the types of metabolic substrates (eg, fat, carbohydrate, and/or protein) being used for energy synthesis.^2,12,13,17 Whether there is a whole-body metabolic response after an SC is currently unknown. Although it is recommended that patients eat a well-balanced diet after injury,¹⁴ empirical evidence supporting this recommendation is lacking. It is imperative to understand the whole-body metabolic response to SC to determine whether nutritional intervention may have a meaningful impact on SC recovery.

At the cellular level, energy systems in the brain are similar to those in skeletal muscle where adenosine triphosphate is created through aerobic and anaerobic metabolism.⁵ Nerve cells in particular rely heavily on glycolysis, which converts glucose into adenosine triphosphate.^1,5,23 The brain also utilizes lactate (a byproduct of anaerobic glycolysis) and ketone bodies (created by fatty acid oxidation in the liver) as fuel sources in the absence of carbohydrates during activities of daily living. Lactate and ketone bodies are used at increased rates during growth and development,¹⁹ intense physical activity,³¹ and injury.²⁷

Following SC, glycolysis in the brain initially increases in an effort to restore homeostasis via both aerobic and anaerobic pathways. Glycolysis subsequently decreases as the demand for glucose outpaces the supply. This impaired glycolytic state necessitates alternate fuel sources.^2,12 There is a dearth of literature addressing the association between brain and whole-body metabolism after SC; however, whole-body energy regulation after moderate or severe traumatic brain injury (TBI) has been investigated.^{4,15,24,32,33} Specifically, resting metabolic rate (RMR) has been measured at up to 200% of average predicted values in the initial days after moderate to severe TBI and remains elevated (116% to 200%) for several weeks.^{4,15,24,32,33} Concordantly, higher concentrations of peripherally measured lactate and ketone bodies have been documented within 9 days after moderate to severe TBI.^13,24 Similar to the brain, this systemically increased availability of nonglucose substrates (eg, lactate and ketone bodies) may represent a whole-body response to an increased overall demand for energy that is not met by glycolysis alone.

The purpose of this study was to examine whether SC alters whole-body energy expenditure in relation to energy intake during recovery. We also sought to investigate potential sex-based differences associated with clinical and energy-related recovery provided that biologic sex may influence clinical recovery from SC.²⁶ We hypothesized that whole-body energy expenditure and dietary energy consumption would increase during the initial days after SC and that these values would normalize throughout recovery. Similarly, we hypothesized that changes in RMR would relate to decreased carbohydrate fuel utilization and that a greater magnitude of change in energy expenditure would relate to longer time to symptom resolution.

Methods

Participants

This study was approved by the university’s institutional review board. Student-athletes between 18 and 29 years of age were recruited to participate in the current study. For the purpose of this study, the definition of concussion was consistent with that provided by the Concussion in Sport group at the time of diagnosis.^21,22 Concussed participants were referred to the research team by their respective certified athletic trainer after a diagnosed injury. Injured participants reported for their first assessment within 72 hours of their diagnosis. Healthy control participants were matched by age, sex, height, weight, and according to their sport/habitual physical activity (eg, varsity athletic team) when possible. All participants provided informed consent prior to study participation. Participants were excluded if they self-reported current treatment for an acute musculoskeletal injury (eg, fracture), a diagnosis of any pathology known to affect metabolism (eg, thyroid dysfunction), or if they had sustained a separate brain injury within 6 months of their first assessment.

Outcome Measures

A VMax Encore Metabolic Cart (CareFusion) was used to measure the exchange of oxygen and carbon dioxide gases (indirect calorimetry).⁵ The VMax Metabolic Cart is a valid measure of RMR when compared with an industry standard and has shown small intrapatient changes in day-to-day measurement (coefficient of variation, 8.4%).⁹ Both RMR (kcal/d) and fuel utilization were measured through indirect calorimetry. RMR was normalized to body mass (RMR/kg) prior to our analyses. The estimated proportions of carbohydrate utilization (%CHO) were calculated using the following equation¹⁸:

$$\%CHO=[(RER-0.71)÷0.29]\times 100$$

where the respiratory exchange ratio (RER) is the relationship between the volumes of expired carbon dioxide and inspired oxygen and is used to approximate the proportion of macronutrients oxidized for energy synthesis.⁵

Total daily energy expenditure (TEE [kcal/d]) was calculated by multiplying RMR by an analogous physical activity level correction factor and was normalized to body mass (TEE/kg). In the TEE calculation, physical activity correction factors were based on estimated activity levels and sex (Table 1).¹⁶ Participants were given a Fitbit Charge HR or Charge 2 (Fitbit, Inc) and asked to track step counts for each day of assessment and 2 subsequent days. Step counts during these days were averaged to represent each time point’s physical activity level (Table 1).^16,30 The Fitbit Charge HR has good evidence of validity when compared with a handheld step counter (intraclass correlation coefficient [ICC] as high as 0.74; 95% CI, 0.54-0.87) and reliability (ICC ≥ 0.70; 95% CI, 0.46-0.86) for the measurement of walking steps.¹¹ When the Fitbit was not worn (eg, while swimming), participants self-reported what activities were performed, thus elevating the physical activity level correction factor by 1 level (eg, from “very active” to “exceptionally active”). For example, a male control participant who recorded 8000 steps on his Fitbit and participated in 2 hours of football practice would be considered “very active” as opposed to “moderately active,” and his measured RMR would accordingly be multiplied by 1.85 for that day.

Table 1.

Physical activity levels and correction factors for calculating total energy expenditure ^a

		Physical Activity Correction Factor
Physical Activity Level			Percentage Above Resting Metabolic Rate
Step Count	Level	Level	Men	Women
<5000	Sedentary	Sedentary	15	15
5000-7499	Low active	Lightly active	40	35
7500-9999	Somewhat active	Moderately active	50	45
10,000-12,499	Active	Very active	85	70
>12,500	Highly Active	Exceptionally Active	110	100

^aStep counts were used to derive physical activity levels for each participant.³⁰ This physical activity level was then used to estimate the expenditure of energy above and beyond their measured resting metabolic rate.¹⁶ Example: Total energy expenditure for a “Somewhat”/”Moderately” active female was equal to her measured resting metabolic rate multiplied by 1.45.

To assess dietary intake, participants completed 3 diet records by self-reporting all food and beverage intake. Intakes were entered into the MyFitnessPal (MyFitnessPal, Inc) online portal by the study team, and energy consumption (EC [kcal/d]) was estimated as the average number of kilocalories consumed by the participant on the day of each assessment and the subsequent 2 days. EC was also normalized to body mass (EC/kg). MyFitnessPal has been demonstrated to yield similar results to traditional standardized paper-based dietary recalls with regard to estimating overall EC (P > 0.61).²⁹ Energy balance (EBal) was calculated as the ratio between EC and TEE (EC/TEE). A value greater than 1.0 indicated that the participant consumed more energy than he or she expended (energy surplus), and a value less than 1.0 indicated an energy deficit.

For concussed participants, days to reporting symptom-free and days to unrestricted return to play were determined through electronic record review (varsity student-athletes) or by self-report (for 1 nonvarsity student-athlete). Participants were asked to complete the Revised Head Injury Scale (HIS-r) daily by their respective certified athletic trainer. The HIS-r includes 3 symptom-related outcomes: the presence of 22 symptoms, the duration (from brief to constant), and severity (not severe at all to as severe as possible) over the previous 24 hours. The duration and severity of symptoms have been demonstrated to have strong sensitivity (77.5%) and specificity (100%) in recognizing the presence of a clinically diagnosed concussion in collegiate athletes.²⁵

Procedures

All participants reported for their initial assessment within 72 hours of their injury (T₁), 7 days after their initial assessment (T₂), and again 7 days after their second assessment (T₃). Participants who were still experiencing symptoms at their third assessment were asked to report for a fourth assessment (T₄) within 72 hours after reporting symptom-free to their athletic trainer. Assessment time points for our analyses included T₁, T₂, and the final assessment time point (T_F). T_F was the last assessment for individuals after reporting symptom-free to their athletic trainer (T₃ for most and T₄ for 1 participant). Control participants were assessed at 3 time points each, separated by at least 3 days between sessions to allow no overlap in recording EC and step counts.

Participants arrived in a fasting state to each assessment between 06:00 am and 09:00 am. On arrival, participants completed a detailed medical history form that included demographic questions and items specific to their concussion history. Next, participants completed the HIS-r, and then height (T₁ only) and weight were measured. Participants were then instructed to lay supine for 20 to 30 minutes with a clear plastic canopy placed over their head that was connected to the VMax Metabolic Cart for indirect calorimetry assessment. Last, step count and EC were recorded for the day of each assessment and for the subsequent 2 days.

Missing Data

In cases of missing RMR, EC, and %CHO data, imputations were made by calculating person-specific values based on the group average change between time points. For example, after EC was normalized to body mass, the group average difference between assessment time points (eg, –5.01 kcal/d/kg between T₂ and T_F) would be multiplied by that participant’s most recently measured body mass and then added to the most recently measured RMR value in the following equation:

$$GroupAverageChange(RMRperkg)\times LastMeasuredMass+LastMeasuredRMR=Participan{t}^{\prime }sEstimatedRMR$$

TEE and EBal were calculated for these individuals using the imputed RMR and EC values. When physical activity was missing, the median group correction factor was used in the calculation of TEE.

Analyses

To assess group-level RMR/kg, %CHO, and EBal outcomes over time, we performed a 2 × 2 × 3 multivariate analysis of variance consisting of group (concussed vs control) by sex (female vs male) by time (T₁, T₂, T_F) comparisons. Effect sizes were calculated as partial eta-squared (h2p) and were interpreted as small (≤0.06), medium (0.06-0.13), or large (≥0.14).⁸ Post hoc comparisons were performed using the Tukey honestly significant difference test with Cohen d effect sizes, which were interpreted as no effect (<0.2), small effect (0.2-0.49), medium effect (0.5-0.79), or large effect (≥0.8).⁸ As described previously, EBal is a construct of physical activity, TEE, and EC; therefore, significant findings in EBal led to subsequent analyses of these individual components. Physical activity was assessed through χ² tests with Cramer V effect sizes, interpreted as small (0.06-0.16), medium (0.17-0.28), or large (>0.29).¹⁰ TEE and EC were assessed with repeated-measures analyses of variance and subsequent independent t tests (group and sex differences) or paired t tests (differences over time). Finally, Spearman ρ correlations were used to examine the relationships between T₁ to T_F changes in RMR/kg, %CHO, EBal, days to symptom-free, and days to return to full participation. All analyses were performed in SPSS Version 25 (IBM Corp), with statistical significance set a priori with α ≤ 0.05.

Results

A total of 20 concussed participants (9 men, 11 women) and 20 matched control participants were included in the study (Table 2). T_F outcomes were imputed for 1 male concussed participant who completed the first 2 assessments but declined to participate in the final assessment, and 1 male control participant sustained a foot fracture after his second assessment. Women reported symptom-free an average of 4 days sooner than men and returned to unrestricted sport participation 5 days earlier than men (Table 2). At T₁ (mean ± SD, 2.1 ± 0.83 days after injury), no concussed participants had symptom resolution. At T₂, 15 of 20 (75%) reported being symptom-free and 3 (15%) had returned to full sport participation. Three male participants had symptoms through T₃, 2 of whom did not return for T₄. Therefore, T₃ outcomes for the 2 who did not return were used as T_F in our analyses. Specific dates of return to unrestricted sport participation were missing for 2 female participants, but only 1 was missing her date of symptom resolution. The remaining 16 participants all reported being symptom-free at T_F and 15 had returned to full participation.

Table 2.

Participant demographics^a

	Concussed Women (n = 11)	Concussed Men (n = 9)	All Concussed (n = 20)	Control Women (n = 11)	Control Men (n = 9)	All Controls (n = 20)
Age, y	19.0 ± 1.10	19.7 ± 1.00	19.3 ± 1.08	19.9 ± 1.30	21.8 ± 2.64	20.8 ± 2.17
Height, m	1.70 ± 0.096	1.85 ± 0.076	1.77 ± 0.112	1.72 ± 0.083	1.82 ± 0.083	1.77 ± 0.096
Weight at T1, kg	65.7 ± 10.21	96.5 ± 24.13	79.6 ± 23.37	69.1 ± 9.69	97.6 ± 26.14	81.9 ± 23.45
Weight at T2, kg	65.6 ± 10.26	96.9 ± 24.27	79.7 ± 23.65	69.9 ± 9.75	97.3 ± 26.19	82.2 ± 23.14
Weight at TF, kg	65.6 ± 10.16	96.7 ± 24.31	79.6 ± 23.56	69.8 ± 9.83	98.0 ± 26.93	82.5 ± 23.75
BMI at T1, kg/m2	22.5 ± 1.66	28.0 ± 5.20	25.0 ± 4.55	23.3 ± 1.98	29.1 ± 5.93	25.9 ± 5.07
BMI at T2, kg/m2	22.5 ± 1.62	28.1 ± 5.17	25.0 ± 4.57	23.6 ± 2.01	29.0 ± 5.96	26.0 ± 4.98
BMI at TF, kg/m2	22.5 ± 1.58	28.0 ± 5.22	25.0 ± 4.55	23.6 ± 1.93	29.0 ± 5.67	26.0 ± 4.81
Concussion history	1 (0-3)	1 (0-6)	1 (0-6)	0 (0-3)	0 (0-2)	0 (0-3)
Days to reporting symptom-free	6 (3-10)	10 (4-29)	6 (3-29)	—	—	—
Days to full return to play	11 (7-16)	16 (10-42)	14 (7-42)	—	—	—

BMI, body mass index; T₁, first assessment time point; T₂, second assessment time point; T_F, final assessment time point.

^aAge, height, weight, and BMI are presented as means ± SDs. The remaining demographic variables are presented as medians (full range).

We did not observe any significant differences in RMR/kg or %CHO over time, between groups, or regarding the interaction of sex and group (Table 3). EBal was different between groups over time (P < 0.003; h2p = 0.14) and is presented in Figure 1. On average, the concussed group was less physically active at the first 2 assessment time points (Figure 2), and TEE/kg was lower in the concussed group compared with the control group at T₁ (Table 3). Additionally, TEE/kg significantly increased over time in the concussed group from T₁ to T₂, and from T₁ to T_F (Table 3). No significant changes over time in TEE/kg were observed in the control group. Energy consumption did not change over time or differ between groups across time points (F₍₂₎ ≤ 2.554; P > 0.08; h2p ≤ 0.066). Changes over time for RMR/kg, %CHO, and EBal in the concussed group overall were not significantly correlated with days to symptom-free or days to unrestricted return to play (P > 0.06); however, in men only, an increase in %CHO utilization over time was significantly correlated with more days to reporting symptom-free and days to return to play (Table 3).

Table 3.

Energy expenditure, consumption, and balance over time^a

	Concussed Women (n = 11)	Concussed Men (n = 9)	All Concussed (n = 20)	Control Women (n = 11)	Control Men (n = 9)	All Controls (n = 20)
RMR per kg at T1	14.4 ± 1.98	14.3 ± 1.63	14.4 ± 1.79	14.5 ± 1.11	14.5 ± 1.80	14.5 ± 1.42
RMR per kg at T2	14.7 ± 1.60	14.1 ± 1.81	14.4 ± 1.66	15.1 ± 1.07	15.1 ± 2.33	15.1 ± 1.70
RMR per kg at TF	14.6 ± 2.12	14.5 ± 1.70	14.6 ± 1.89	14.7 ± 1.35	14.3 ± 2.05	14.5 ± 1.67
%CHO at T1	46.0 ± 22.1	31.0 ± 14.9	39.2 ± 20.2	46.7 ± 13.7	30.5 ± 21.8	39.4 ± 19.2
%CHO at T2	43.9 ± 19.0	44.9 ± 21.1	44.4 ± 19.5	40.8 ± 14.1	41.6 ± 28.7	41.2 ± 21.2
%CHO at TF	45.1 ± 21.8	41.4 ± 18.8	43.4 ± 20.0	42.4 ± 14.7	44.8 ± 21.4	43.5 ± 17.6
%CHO change from T1 to TFb,c	−0.91 ± 24.5	10.4 ± 15.4	4.19 ± 21.2	−4.26 ± 17.0	17.3 ± 21.4	5.44 ± 21.6
Number who increased %CHO from T1 to TFc	4/11 (36.4%)	5/8 (62.5%)	9/19 (47.4%)	4/11 (36.4%)	6/8 (75.0%)	10/19 (56.2%)
TEE per kg at T1d	21.0 ± 1.85	23.6 ± 4.55	22.2 ± 3.49	26.4 ± 3.32	27.2 ± 4.62	26.8 ± 3.87
TEE per kg at T2e	24.2 ± 2.66	24.7 ± 5.88	24.4 ± 4.28	27.3 ± 4.50	28.0 ± 5.42	27.6 ± 4.81
TEE per kg at TFe	25.0 ± 5.62	29.1 ± 5.45	26.8 ± 5.78	26.0 ± 4.03	26.4 ± 5.01	26.2 ± 4.37
EC per kg at T1	28.3 ± 7.65	26.6 ± 10.94	27.6 ± 9.05	24.1 ± 4.83	28.0 ± 6.71	25.8 ± 5.93
EC per kg at T2	26.8 ± 9.01	28.2 ± 11.23	27.4 ± 9.81	22.9 ± 6.20	24.3 ± 6.58	23.5 ± 6.24
EC per kg at TF	26.1 ± 6.02	23.2 ± 6.81	24.8 ± 6.39	24.8 ± 6.87	26.7 ± 6.58	25.6 ± 6.63
EBal at T1	1.34 ± 0.32	1.15 ± 0.43	1.26 ± 0.38	0.92 ± 0.19	1.07 ± 0.40	0.99 ± 0.30
EBal at T2	1.10 ± 0.31	1.18 ± 0.47	1.14 ± 0.38	0.85 ± 0.24	0.91 ± 0.17	0.87 ± 0.21
EBal at TF	1.08 ± 0.30	0.81 ± 0.22	0.96 ± 0.30	0.95 ± 0.22	1.09 ± 0.24	1.01 ± 0.24

%CHO, percentage carbohydrate utilization; EBal, energy balance; EC per kg, energy consumption per kilogram of body mass; RMR, resting metabolic rate; T₁, first assessment time point; T₂, second assessment time point; T_F, final assessment time point; TEE per kg, total energy expenditure per kilogram of body mass.

^aThere were no observed statistically significant differences in normalized RMR (RMR per kg) or the percentage of RMR coming from carbohydrate metabolism (%CHO) over time, between groups, or regarding the interaction of sex and group. Data presented as means ± SDs unless otherwise indicated.

^bIn concussed men only, increased %CHO from T₁ to T_F was significantly correlated with days to reporting symptom-free (ρ = 0.735; P = 0.04; n = 8) and days to return to play (ρ = 0.829; P = 0.02; n = 7).

^cImputed data for 1 concussed male patient and 1 control male participant were not included in the analysis of the individuals who increased %CHO from T₁ to T_F.

^dSignificantly lower in control participants compared with concussed participants: t(38) = −3.938; P < 0.01; d = 1.25 [0.57-1.93].

^eSignificantly higher than T₁ in the concussed group only: T₂: t(19) = −3.269, P < 0.01, d = 0.56 [−0.07 to 1.20]; and T_F: t(19) = −3.723, P < 0.01, d = 0.96 [0.31-1.62]. Energy consumption did not significantly change over time and was not different between groups: F₍₂₎ ≤ 2.554; P > 0.08; h2p ≤ 0.066.

Figure 1.

Energy balance over time. Total n = 40. There was a significant reduction in EBal from T₁ to T_F in the concussed group (P < 0.01; d = 0.89 [0.24-1.54]), and the concussed and control groups were different at both T₁ (P = 0.016; d = 0.90 [0.25-1.55]) and T₂ (P = 0.010; d = 1.26 [0.58-1.94]). Conc, concussed group (n = 20); Ctrl, control group (n = 20); EBal, energy balance; EC, energy consumption; F, female participants; M, male participants; T₁, first assessment time point; T₂, second assessment time point; T_F, final assessment time point; TEE, total energy expenditure.

Figure 2.

Physical activity levels at each assessment time point. (a) Physical activity levels at first assessment time point, T₁. Concussed and control groups were significantly different: χ²₍₅₎ = 16.640; P = 0.01; V = 0.65. (b) Physical activity levels at second assessment time point, T₂. Concussed and control groups were significantly different: χ²₍₅₎ = 11.800; P = 0.04; V = 0.54. (c) Physical activity levels at the final assessment time point, T_F. Concussed and control groups were not significantly different: χ²₍₅₎ = 4.956; P = 0.42; V = 0.35.

Discussion

Based on previous studies in patients with moderate or severe TBI, we expected a significant hypermetabolic state as a result of SC. However, a hypermetabolic state was not observed. Previous studies that reported an increase in whole-body metabolism included participants with more significant brain injuries and potentially confounding comorbidities.^2,7,13,17,24 We may have also missed the metabolic window to recognize the previously described hypermetabolic state in our participants. Another possible explanation could be the influence of low levels of physical activity after injury in the SC group. Physical activity was different between groups in the first 2 assessment time points and increased over time in the concussed group, despite no significant group or group by time differences in RMR. Residual increases in energy expenditure occur after bouts of physical activity; and therefore, physical inactivity could be expected to result in relatively decreased RMR.³ The lack of group differences regarding RMR in this study could be related to the off-setting influences of injury (elevated RMR) and lower levels of physical activity (decreased RMR).

Lower physical activity levels in the SC group combined with slightly elevated EC during T₁ and T₂ appeared to drive the differences in EBal between groups and over time. Acutely, average EBal was almost 130% of TEE in the SC group, and this value returned to relative caloric balance after participants reported symptom-free. There is a tendency for athletes to underreport the amount of energy they consume by an average of 19% across studies, and we therefore may have underestimated our sample’s EBal.⁶ Even so, we did observe isocaloric EBal at the first and final assessment time points in our control group. From this, we can infer that any present reporting error was systematic within our study design and the resulting bias did not affect our change-estimates differently for each group. In sum, decreased physical activity likely contributed to the disparity in TEE, and when combined with increased EC, we are confident that EBal was significantly affected by SC. Currently, it is unknown whether positive EBal in collegiate student-athletes with SC has a healing effect and whether it is a response to changes in physiology and/or behavior. Thus, further observational investigations of dietary intake are warranted after SC, and dietary interventions (eg, hypercaloric diets) should be explored.

This study has limitations that merit discussion. First, we did not obtain preinjury baseline assessments of energy expenditure, physical activity, or dietary intake; however, we matched concussed participants to healthy control participants based on age, sex, height, weight, and sport. Moreover, healthy controls were serially measured like the concussed group to have meaningful comparators over time in lieu of preinjury assessments. Second, total ambulation via Fitbit and self-reported activities was used to gauge the amount and intensity of physical activity, thereby limiting accuracy of activity-related energy expenditure. Additionally, estimated EC may have been affected by the ability of concussed participants to keep track of the type and amounts of foods consumed. Diet records, even when used in healthy and alert athletes, are associated with underreporting of dietary intake and have the potential to influence eating behavior.^6,28 Because of financial restrictions and considerations of practicality for our participant sample, we did not attempt to validate self-reported diets using physiological biomarkers of intake. Therefore, we were unable to determine the direction or magnitude of self-reporting errors or the presence of bias in this study. In an effort to limit these sources of error, we encouraged participants to report their food and beverage intakes consistently and truthfully for 3 consecutive days, and a member of the research team was regularly available to address questions or concerns.^6,20 Finally, we were unable to perform an a priori power calculation due to the lack of available data. As such, our statistical power is likely inadequate to answer our research questions.

Conclusion

Whole-body metabolic rate (RMR/kg) was not affected by SC. Concussed participants overconsumed energy (hypercaloric state) compared with their relative energy expenditure within the first 10 days of injury, followed by a return to an isocaloric state after reporting symptom-free. Female participants with SC recovered more quickly than their male counterparts, though only men with SC exhibited changes in %CHO over time that were strongly correlated with longer recovery times. The relevance of this finding remains unknown as control male participants exhibited a similar pattern of %CHO change over time. Taken together, these data are suggestive that SC may have unique effects on energy consumption versus utilization, and interventions around diet and physical activity may be helpful avenues to explore to improve clinical recovery.