**1. Introduction**

From Olympians to recreational exercisers, athletes of all levels face the same questions—what should I eat before exercise, and how does it a ffect my training? Despite being relevant to anyone performing exercise, many questions relating to the e ffects of nutritional intake on endurance training responses and adaptations remain unanswered.

The duration and intensity of exercise are the most important factors influencing the adaptive response to endurance training [1]. However, strategies altering nutrient availability before and during exercise can also impact training adaptations by modulating the exercise stimulus and/or cellular responses to the exercise-induced perturbations [2]. Specific strategies to alter nutrient availability can include exercising in the overnight-fasted state, restricting carbohydrate (CHO) ingestion between training sessions, and increasing CHO ingestion before or during exercise [3]. Although performance may be improved following pre-exercise CHO ingestion [4,5], exercise undertaken with reduced availability of CHO can increase the activation of key signaling proteins compared with exercise performed with high CHO availability [6], potentially influencing longer-term training adaptations.

Among the intracellular signals comprising the endurance training response are mechanical stretch, reactive oxygen and nitrogen species (RONS), calcium flux, AMP:ATP ratio, and the availability of endogenous CHO and free fatty acids (FFA) [7,8]. These signals are a ffected by both the duration and intensity of an exercise session, and by the pre-exercise nutrition choices of an athlete (i.e., the size, type, and timing of the pre-exercise meal(s), Figure 1). Although some lines of evidence sugges<sup>t</sup> ingesting CHO before exercise can negatively influence endurance training adaptations, contrasting findings have been reported. For example, ingesting CHO has decreased [9], increased [10], or had no effect [11] on the activity of the 5 AMP-activated protein kinase (AMPK) following exercise. Similarly, training-induced improvements in maximal oxygen consumption (VO2max) have been reported to increase [12], decrease [13], or remain unchanged [14] following 4–6 weeks of CHO-fed compared with fasted-state training. These contrasting findings can be a source of confusion and may explain why the beliefs and practices relating to the role and influence of pre-exercise nutrition vary so widely among coaches and athletes [15,16]. Accordingly, the purpose of this review is to highlight the current knowledge of the influence of pre-exercise nutrition ingestion on the metabolic, physiological, and performance responses to endurance training. We also highlight areas for practitioners where evidence is lacking, particularly regarding trained athletes, and sugges<sup>t</sup> directions for future research.

**Figure 1.** Schematic of areas where pre-exercise nutrition has the potential to impact the adaptive responses to endurance training. Green arrows sugges<sup>t</sup> the potential to increase or augmen<sup>t</sup> specific signaling, and red dashed arrows sugges<sup>t</sup> the potential to decrease or impair specific signaling. Abbreviations: AMPK, AMP-activated protein kinase; CaMK, calcium/calmodulin-stimulated protein kinase; CHO, carbohydrate; FFA, free fatty acids; LCHF, low-CHO high-fat; MAPK, mitogen-activated protein kinase; VO2max, maximal oxygen consumption.

#### **2. Acute Responses to Pre-Exercise Nutrition Intake**

The vast majority of pre-exercise nutrition interventions have been conducted in an acute context. Although acute responses to training do not always correspond with long-term adaptations [17,18], the accumulation over time of transient, exercise-induced changes in gene expression are thought to be the driving factor behind many adaptations to training [19]. Therefore, it is relevant to consider the acute e ffects of pre-exercise nutrition in addition to the longer-term adaptations.

## *2.1. Metabolism and Substrate Oxidation*

The liver plays a key role in metabolic regulation during extended exercise [20]. Despite the ~40% reduction in liver glycogen following an overnight fast [21], blood glucose concentration can be maintained at normal levels during exercise due to increased gluconeogenesis and/or decreased

utilization of glucose in skeletal muscle [22,23]. However, fatigue during extended exercise is often associated with reduced blood glucose concentrations [24], supporting a critical role for liver glycogen in achieving optimal performance during extended exercise.

Exercising in the fasted-state generally allows higher levels of fat oxidation than exercise performed in the CHO-fed state during low-to-moderate intensity exercise [25] and can increase the relative intensity where maximal fat oxidation occurs [26]. Ingesting CHO before exercise increases plasma glucose and insulin levels, leading to a reduction in hepatic glucose output and an increase in skeletal muscle glucose uptake during exercise [27]. This can lower fat oxidation by decreasing plasma FFA availability via insulin-mediated inhibition of lipolysis [28], and also by inhibiting fat oxidation within the muscle due to an increased glycolytic flux [29]. Intramuscular triglycerides (IMTG) provide a key substrate for fat oxidation, primarily during exercise in the fasted state [30,31], although their use declines as the duration of exercise extends, while the oxidation of plasma FFA increases [32]. Up to 6 h may be required following a CHO-rich meal for substrate oxidation and glucose homeostasis to return to levels observed during fasted-state exercise [33].

In contrast with exercise performed in the overnight-fasted state, which lowers hepatic but not muscle glycogen [34], restricting CHO between training sessions allows exercise to be undertaken with reduced muscle glycogen concentrations [35]. During exercise with low muscle glycogen there is an increase in the oxidation of fat [36,37] and amino acids [38,39], and a reduction in muscle glycogen breakdown [36,40,41]. During exercise undertaken with normal muscle glycogen levels, muscle glycogen breakdown is similar between fed and fasted-state exercise [31,42–44] and may be reduced when ingesting CHO during exercise [45].

The majority of research looking at fat oxidation has compared CHO to a placebo, but the use of pre-exercise protein ingestion represents an interesting and under-researched area. Consumption of protein before and during steady-state exercise did not affect FFA availability or whole body fat oxidation compared with fasted-state exercise commenced with normal [46] or lowered [47] muscle glycogen concentration, despite elevated insulin levels. This may be related to the increases in catecholamine levels during exercise, which are an important determinant of the adipose tissue lipolytic rate and can override the inhibition by insulin [48]. Although protein ingestion before exercising in a low-glycogen state has no effect on rates of muscle protein synthesis, it is plausible that it could reduce muscle protein breakdown during exercise [49]. It also appears possible that pre-exercise protein ingestion increases amino acid oxidation during exercise [49], but further quantification of its influence is needed.

To compare the influence of pre-exercise CHO ingestion, muscle glycogen levels, and glycemic index on substrate oxidation and AMPK activity, we pooled the results of 125 studies (available as supplementary online files) that included the relevant intervention groups (Figures 2–7). Together, these studies included 1245 subjects (12.8% female), with an average age, BMI, and VO2max of 25.4 ± 3.1 years, 23.2 ± 1.4 kg m2, and 56.7 ± 8.2 mL kg−<sup>1</sup> min−1. Linear correlation analysis was used to calculate the correlation coe fficient between variables, according to Pearson's product moment (r) using R statistical software. Pooled data are reported as mean ± SD, with the level of statistical significance set at *p* < 0.05.

#### 2.1.1. E ffect of Exercise Duration

The respiratory exchange ratio (RER—a measure of substrate oxidation) decreases with exercise duration, indicating an increasing reliance on fat oxidation as the duration of exercise extends [50]. Di fferences in RER between exercising in the fed vs. fasted state and following low vs. high glycemic index CHO remain largely similar throughout exercise, while the di fferences in RER between high and low starting muscle glycogen decrease as exercise duration extends (Figure 2). The latter could presumably be related to the greater utilization of muscle glycogen during exercise undertaken with higher levels of glycogen, leading to more similar levels during the later stages of exercise. This idea is supported by the pooled data, which show a strong correlation (r = 0.89, *p* < 0.001) between the di fferences in pre-exercise glycogen levels and di fferences in RER during exercise (Figure 3).

**Figure 2.** Substrate oxidation in relation to exercise duration for studies reporting respiratory exchange ratio (RER) at multiple time points comparing overnight-fasted and/or CHO-fed exercise with normal muscle glycogen levels (**A**), exercise undertaken with high (471 ± 208 mmol kg−<sup>1</sup> dry mass) and low (232 ± 112 mmol kg−<sup>1</sup> dry mass) muscle glycogen levels (**B**), and following high (82 ± 10) and low (36 ± 9) glycemic index meals (**C**). Shaded areas represent 95% confidence intervals. Data were obtained by pooling results from 60 studies (see supplementary files for references).

**Figure 3.** Correlation between differences in respiratory exchange ratio (RER) during exercise and differences in pre-exercise glycogen levels. Shaded area represents 95% confidence intervals. Data were obtained by pooling results from 13 studies that manipulated glycogen levels and reported RER for high- and low-glycogen trials (see supplementary files for references). DM = dry mass.
