Protein intake during training sessions has no effect on performance and recovery during a strenuous training camp for elite cyclists

To evaluate the acute effect of the intervention, a test protocol was performed in the morning at Day one (D1pre) and repeated after four (short distance group) or five (long distance group) hours of cycling (D1post). The test protocol was repeated at day six in the morning (D6pre). Breakfast and time points for testing at Day six were the same as at Day one of the intervention. The protocol consisted of a standardized warm-up prior to a 10-s peak power test followed by a 3-min recovery period and a 5-min all-out performance test.

Blood samples were collected before (morning sample) and immediately after the performance tests at Day 1 and Day 6 as well as in the morning of day 3, 5, and 7, and analyzed for markers of muscle damage (creatine kinase (CK), lactate dehydrogenase (LDH), and myoglobin) and cortisol. Saliva samples were collected in the morning after overnight fasting at Day 1, 5, 6, 7, immediately after exercise, and 1 h after the training session on Day 6 to be analyzed for immunoglobulin A (IgA).

The study complied with the Declaration of Helsinki and was approved by the local ethics committees (1-10-72-558-12). All subjects gave their informed consent to participate prior to the experiments.

During the intervention period, each cyclist received one bottle (750 ml) containing the intervention beverage for every hour of exercise performed. The PRO-CHO group ingested a beverage containing 0.2 g of protein/kg/h (Whey protein hydrolysate with a degree of hydrolysis between 23–29 %, Arla Foods Ingredients Group P/S, Viby, Denmark) and 1 g of CHO/kg/h (Maxim Energy Drink, Maxim International, Ishoej, Denmark). The CHO group consumed an energy-matched carbohydrate beverage containing 1.2 g of CHO/kg/h (Maxim Energy Drink, Maxim International, Ishoej, Denmark). The carbohydrate content in both beverages consisted of maltodextrin and fructose in the ratio 2:1. We chose an isocaloric control beverage rather than a beverage matched for carbohydrate to isolate the effect of the addition of protein rather than an extra amount of energy.

Residual fluid left in used bottles was less than 300 ml in total per training day except for one subject in the PRO-CHO group who was not able to drink the total beverage volume on three of the six days (residual fluid remaining was 400 out of 4500 ml, 600 out of 3750 ml, and 1875 out of 3750 ml, respectively).

Both groups consumed a recovery beverage immediately after exercise each day containing 18 g of protein and 69 g of carbohydrate corresponding to ~0.25 g protein/kg and 1 g carbohydrate/kg (similar sources of protein and carbohydrate as during cycling). Nothing else was ingested, except water, within the two hours before and the one hour after each training session each day. Neither the subjects nor the testing staff was informed about the content of the beverages. Beverages were prepared by staff members not present during the training and performance testing. At D1, the recovery beverage was ingested after the post-exercise testing (D1post).

On Day 6, individual weight changes during the 6 h training session were determined to check hydration status. The athletes were weighed immediately before and after the training session in minimal clothing. The individual bottles were weighed before and after the 6 h training session. On average the athletes were provided with 4.33 L in total during the training session (0.72 L/h). One subject did not consume 0.35 L of the provided beverages during the training session and two subjects consumed additional 0.50 and 0.75 L plain water. Weight data from these three subjects were adjusted in the evaluation of the effect of 0.72 L/h on hydration status (weight change).

The training regime was standardized and controlled by the National U23 coach in cycling. The subjects were divided in two groups depending on their training status and history (long: 28.6 h/6 days and short 24.9 h/6 days). The cyclists in the matched pairs were within the same training groups and thereby an equal number of long and short distance subjects were represented in PRO-CHO and CHO respectively. The long distance group biked ~5, 5, 1, 6, 5, and 6 h during the 6 training days (plus standardized warm-up and tests) and the short distance group biked 4, 4, 1, 4, 5, and 6 h (plus standardized warm-up and tests). The training program within both the short and long distance group consisted of a mix of distance training, interval training, mountain climbing, and an individual time trial. The total number of training hours and training distance were identical in the two intervention groups since cyclists in the matched pairs were cycling together.

To avoid any dietary bias, subjects followed a predetermined energy balanced diet plan throughout the week. Prior to the intervention week, each cyclist had an individual meeting with a dietitian where the importance of the diet control was emphasized. Furthermore, food preferences were noted and the subjects were asked if they were suffering from food allergies. Afterwards, an individual diet plan was prepared based on the weight, training load, and preferences using the online software madlogvita (http://www.madlogvita.dk↗). Total daily energy expenditure (TEE) was estimated for each training day during the intervention and the diet plans varied between days, depending on training schedule. Total daily energy expenditure was estimated based on the following equation: TEE (kJ/day) = RMR x PAL + EX, where RMR is the estimated resting metabolic rate [28], PAL is estimated physical activity level in non-training hours, which was set at 1.5 corresponding to a sedentary lifestyle [28], and EX was the estimated energy used during training sessions (number of training hours × 200 W × 3.6 × (100/25)), assuming an exercise effectiveness of 25 % and an average power output of 200 W. The food was served as a buffet and the participants weighed all the food items in accordance with their individual diet plan. A dietitian was available at the buffet each morning and evening. In addition, participants could always contact and meet with the dietitian if they had a question about their individual diet plans. The subjects were not allowed to consume dietary supplements, sports products, or any food item apart from the items in their individual diet plans.

The average distribution of macronutrients in the diet (excl. intervention beverages) was 8 g of carbohydrate/kg/day (~62 % of energy (E%)) and 1.7 g protein/kg/day. The 18 g protein in the recovery beverage ingested by both groups was included in the 1.7 g protein/kg/day, whereas the protein content in the intervention beverages ingested by PRO-CHO during cycling was in addition to the controlled diet. Fat was supplemented to meet each individual’s energy need (~22–24 E%). The daily intake of protein in CHO corresponded to protein recommendations for elite endurance athletes [29]. The protein intake in PRO-CHO, including the intervention beverages, was 2.6 g/kg/day. Total daily carbohydrate intake including beverages was 14.6 g/kg/day in CHO and 13.6 g/kg/day in PRO-CHO. The energy content of the diet was adjusted (±125 kcal/day) to reestablish energy balance if morning weight had changed more than 1 kg body weight and a bioimpedance measurement did not indicate that the weight change was due to change in hydration. Average energy intake was 24.4 MJ/day in long (29 h training) and 22.0 MJ/day in short (25 h training), including 8.4 MJ/day (long) and 7.2 MJ/day (short) from intervention beverages. The food and nutrient composition of the breakfast before the test protocol at D1 and D6 was the same to ensure standardization of the subjects.

The subjects completed a VO_2max test on a stationary bike ergometer (SRM) one or two weeks before the training camp. After a 15-min warm-up, subjects were instructed to work all-out for 5 min during which they were allowed to change load and cadence. VO₂ was continuously measured through a mask connected to AMIS software (AMIS 2001, Innovision, Odense, Denmark). Every 30 s the average respiratory variables were registered. Heart rate (HR) was continuously recorded using POLAR RS800 or RS800CX. Lactate was measured 1 and 3 min after the test. To ensure that VO_2max had been obtained, two of the following four criteria had to be met [30]: 1) VCO₂/VO₂ > 1.10, 2) HR was within ±5 beats/min of maximal HR (HR_max, based on earlier test results), 3) Plateau in oxygen consumption, or 4) Accumulation of lactate (>8 mmol/L).

The 10-s peak power tests and the 5 min all-out tests were performed on an ergometer bike (Monark Ergomedic model 894E, Monark Exercise AB, Sweden). Within the two weeks before the training camp, the cyclists were familiarized to the test protocol and equipment used during the intervention period. The protocol consisted of a 10-s Peak Power Test, 3 min recovery, followed by a 5-min all-out performance test. The pre-intervention performance test was carried out identically to those during the training camp intervention (same bikes and individual bike settings). The results from the pre-intervention 5-min all-out performance test were used to match the cyclists before block randomization. Blood samples were collected 1 and 3 min after completion of the performance test and analyzed for lactate.

Venous blood samples were immediately centrifuged and frozen. Samples were analyzed for markers of muscle damage (CK, LDH, and myoglobin) and for cortisol at the Department of Bioscience, Aarhus University Hospital. Coefficient of variation (CV) for CK, LDH, myoglobin, and cortisol determination was 2.8–3.9 %, 3.0–4.0 %, 6.8–9.0 %, and 5.9–6.0 %, respectively, depending on the concentration level.

Capillary blood was obtained by finger sticks and immediately analyzed for lactate by an enzymatic lactate analyzer (Yellow Springs Instruments lactate analyzer model 23 L).

Saliva samples were analyzed for IgA with a Salivary Secretory IgA Enzyme Immunoassay kit (Salimetrics Europe, Ltd., Item No. 1–1602, Suffolk, UK) using a BioRad Model 550 Microplate Reader and Microplate Manager version 5.2.1. Morning saliva samples were collected during a 3-min period and post-exercise samples during 2 min. The analytic results were corrected for differences in saliva flow rate by multiplying the results by the saliva produced during the collection period (flow rate: ml min⁻¹).

Data were log-transformed when appropriate based on an inspection of the standardized residuals (CK, LDH, cortisol). Overall, performance data, CK, LDH, myoglobin, cortisol, IgA, and weight were analyzed using multivariate repeated measurements ANOVA with treatment (PRO-CHO and CHO) and time interactions, followed by Student-Newman-Keuls Post-hoc test if significance was observed. The unequal standard deviations and correlations in the two groups were taken into account in all the analyses by letting the standard deviations and correlations vary between groups. In more detail, the performance data was tested separately for an acute effect (D1pre vs D1post) and for a long-term effect (D1pre vs D6pre). Data for body weight and IgA was analyzed for an acute effect on D6 (before and after training the training session). The CK, LDH, myoglobin, cortisol, IgA, and body weight data were tested for a long-term effect (D1 to D7), both for interaction and time effect as well as separated one-way ANOVA analysis for the PRO-CHO and the CHO groups if interaction effect was significant.

The data was analyzed using Stata version 12.1. P values set at <0.05 were taken to indicate statistical significance. Data is reported as means ± standard error of the mean (SE) if nothing else is stated.

The average power during the 5-min all-out performance test at D1-pre was 426 ± 11 W and 428 ± 11 W in CHO and PRO-CHO, respectively (p = 0.88) (Fig. 3). The average power output was lower (p < 0.01) after training at D1 (D1-post) in both CHO and PRO-CHO (409 ± 8 W and 404 ± 11 W) as well as in the morning of D6 (410 ± 11 W and 404 ± 11 W) (Fig. 3).

Plasma myoglobin was below detection level (<53 μg L⁻¹) in all morning samples except for one subject in PRO-CHO, where myoglobin level reached 73 μg L⁻¹ at D3 (data not shown).

In the morning samples, cortisol increased over time (p < 0.001) and was significantly higher at D6 and D7 compared to Day 1, but no significant interaction effect was observed (p = 0.50; Fig. 5b).

Body weight measured after overnight fasting did not change significantly from the first to the last day of the training camp in PRO-CHO (−0.3 ± 0.8 kg) or CHO (−0.6 ± 1.0 kg). During the six-hour training session at D6 the cyclists consumed 0.72 L/h (in total ~4.3 L). Nevertheless, the body weight in PRO-CHO was significantly lower after the training session at D6 (71.5 ± 2.3 kg) compared with immediately before the training session (72.1 ± 1.4 kg, p < 0.05), whereas no significant difference was observed in CHO (71.6 ± 1.3, vs. 71.9 ± 2.5 kg, p = 0.17). However, the change in body weight was not significantly different between groups (−0.8 ± 0.3 % vs. -0.4 ± 0.3 % in PRO-CHO and CHO, respectively, p = 0.41).

Potential positive effects of protein ingestion, such as glycogen sparing, maintaining Krebs cycle intermediates, or reducing central fatigue [31], have led to the hypothesis that protein ingestion during exercise would improve endurance performance. We observed no acute effect of performance on a subsequent 10-s sprint test and the 5-min performance test after consumption of 14 g protein/h during 4 or 5 h cycling compared to consumption of isocaloric carbohydrate beverages. This is in line with several other studies showing no advantage of adding protein to carbohydrate beverages during exercise [3, 5, 6, 32, 33]. Our trial adds support to this general observation by providing new data based on elite endurance cyclists. Studies that have found an acute performance effect of protein supplementation have either been non-isocaloric [9, 7, 10] or have examined performance effects during heat stress [8]. The importance of comparing isocaloric beverages rather than non-isocaloric beverages (matched for carbohydrate content) ingested during exercise on endurance capacity is underlined by results from two studies from the same research group using almost identical protocols [9, 15]. When using an isocaloric control beverage no difference in TTE was observed when partly substituting carbohydrate with protein [15]. In contrast, when effects of beverages matched for carbohydrate content was studied, TTE was improved (29 %, 106 vs 82 min) during cycling when protein was added to the beverage [9]. The latter finding may be a consequence of either a 20 % higher energy intake when protein was added or a suboptimal carbohydrate intake in both groups (~38 g/h). When the carbohydrate intake during endurance exercise is in accordance with the recommendations [1, 2, 29], as in our present trial, no acute beneficial effect on TT [3, 34] or TTE [6, 15, 35–37] of either partial substitution of carbohydrate with protein or adding protein to the carbohydrate supplement is reported.

Even if protein ingestion during exercise does not have an acute beneficial effect on performance when the recommended amount of carbohydrate is ingested, it could have a positive effect on subsequent recovery and performance. Optimizing recovery is especially important for elite cyclists who often have to compete several days in a row. We observed a decline in performance at Day 6 compared to before initiating the intervention, but the decrease in performance was not significantly different between groups as hypothesized. This finding is in line with studies of moderately or well-trained subjects where protein along with carbohydrate ingested during and immediately after exercise compared to carbohydrate beverages show no effect on TTE [15] or TT [38] performed less than 24 h after exercise completion. Nevertheless, in a randomized cross-over trial, 10 well-trained cyclists (66.2 ± 6 mL O₂/kg/min) experienced improved performance (1.8 %) in a TT (~20 k) after 4 h of recovery when the athletes had ingested a PRO-CHO beverage (0.33 g/kg/h + 0.87 g/kg/h) vs isocaloric CHO beverage (1.2 g/kg/h) during a 2.5 h cycling bout followed by a PRO-CHO recovery beverage immediately after the exercise session [11]. The content of the beverages ingested during the initial exercise bout was in accordance with the beverages used in the present study, but the discrepancy in theirs [11] our findings is likely related to either the duration of the recovery period between exercise sessions or the type of test. In the majority of road bike races, an overnight recovery is possible. The prolonged overnight recovery may render any benefits of protein ingestion during exercise inconsequential compared to shorter recovery duration.

The athletes in the present study were supplied 18 g protein/CHO immediately after the training sessions as part of their normal routine, and we aimed to investigate whether additional protein during the training session could prevent the gradual decrease in performance during repeated days of strenuous cycling during a training camp. In contrast to our hypothesis, performance in the 5-min “all-out” test declined similarly in CHO and PRO-CHO during the 6 days of training camp, whereas 10-s sprint performance did not decline. The effect of combined CHO-PRO vs CHO supplementation on cycling performance during eight days of a strenuous mountain bike race in a hot environment has been studied in a previous study [20]. In this study, the PRO-CHO supplemented group completed the race faster than the CHO supplemented group, but in contrast to our study, ingestion of the beverages was not controlled but ingested ad libitum. Furthermore, the daily total energy content was not controlled, and in the CHO group, the body weight was reduced after the intervention period. Therefore, the performance-enhancing effect of PRO-CHO compared to CHO supplementation during repeated days of racing in the previous trial [20] may be related to negative energy balance in the CHO group rather than the addition of protein to the carbohydrate beverage.

In the present trial the diet was energy-balanced, standardized, and controlled. The daily protein intake in the control group (CHO) was 1.7 g/kg, corresponding to recommendations for elite endurance athletes [39–40]. A carbohydrate-rich diet was served to both groups to negate any effect of suboptimal carbohydrate intake on recovery of glycogen stores between training sessions and a recovery-beverage containing both carbohydrate and protein was served immediately after each training session in line with the recommendation and normal practice within this group of elite cyclists. Therefore, we cannot rule out that protein ingestion during exercise would be beneficial on performance under circumstances where the nutrient guidelines are not followed and/or the energy intake is insufficient.

The acute and accumulated reduction in 5-min all-out performance during the training camp reflected the high physical demand of the camp. A time trial test was chosen rather than a time to exhaustion test because time trial tests more closely simulate an actual performance situation and hence are regarded a more valid measure of performance [41]. Nevertheless, the performance in the 10-s sprint test was not significantly affected after either 4–5 h cycling or the accumulated hours of cycling during the training camp. The latter indicates that this test was not sensitive to catch development of fatigue and impairment in recovery. The 10-s peak power test was chosen as peak power is typically obtained after 3–5 s in a 30-s Wingate test [42], and this test allows quick recovery with the impending 5-min performance test in mind.

LDH and myoglobin as markers for muscle damage were not significantly different between the groups. Surprisingly, the results suggest that carbohydrate supplementation rather than protein plus carbohydrate attenuated the increase in plasma concentration of CK as a marker for muscle damage during the strenuous training camp. However, there was a large variation in CK between the subjects and the overall CK values were at the low end. Additionally, it must be emphasized that CK was higher at baseline in the CHO group due to high values in two subjects. Therefore, one should be cautious when drawing conclusions. Overall, the changes in markers for muscle damage were small, which probably is explained by the fact that cycling is primarily based on concentric muscle contractions. In agreement with the latter suggestion, orienteering has a greater proportion of eccentric muscle actions, and in a group of elite orienteers we in an earlier study observed an increase in markers for muscle damage in the group ingesting carbohydrates before and after each of 13 training sessions during a one-week training camp [21]. This increase was attenuated in the group ingesting protein before and a protein-carbohydrate beverage after each training session. Although not all trials are in agreement [20, 40, 43], many previous trials [15, 17, 19, 21, 23, 44] show that protein ingestion immediately after exercise attenuates markers of muscle damage. Therefore, the addition of a protein supplement during exercise may further attenuate the increased rate of muscle damage that is associated with exercise that includes an eccentric component. Future trials should clarify if protein supplementation during exercise has an additive effect for decreasing muscle damage beyond what has been observed when provided immediately after exercise.

Studies have shown that high availability of carbohydrate reduces the immune depression associated with extreme endurance training in athletes [45]. However, little is known about the effect of protein supplementation on immune function in relation to exercise [45]. In the present study, changes in the immune marker salivary IgA did not differ between groups, which may be related to the fact that subjects were in energy balance and the carbohydrate intake was high in both groups.

A drawback of the present study was that we were not allow to obtain muscle biopsies from the subjects since they were all part of or candidates for the Danish National Team in cycling (U23). Analysis of muscle biopsies could have given information about if the adaptation to training stimuli differed between the intervention groups.

This field study was performed at a training camp with a temporarily established laboratory. Since the time frame for testing all athletes in the beginning and end of training camp was very short, it was only possible to perform a short-term performance test. We assumed that extra protein ingested during training could reduce muscle damage and this would be reflected in a short-term test. Nevertheless, we cannot rule out that ingestion of extra carbohydrate during cycling in CHO compared to PRO-CHO actually would be beneficial for performance in a more long-term endurance test due to greater glucose availability.

It is debatable whether our results are transferable to other groups of elite endurance athletes than cyclists. Cycling is dominated by concentric muscle contraction and probably only a minimum of muscle damage is induced, even though the cyclists increased their training load during the training camp. We cannot exclude that other types of athletes (runners and triathletes) may benefit from ingestion of protein during exercise when following a strenuous training regime, particularly those that use predominantly eccentric muscle contractions (such as downhill running and resistance training).

A recovery beverage was provided after each training session with a dose of protein corresponding approximately to the amount of high-quality protein that appears to maximize the acute anabolic response in skeletal muscle [46]. We hypothesize that the recovery beverage may have diminished the potential positive effect of ingestion of protein during exercise. Therefore, it is possible that we would have observed a positive effect in PRO-CHO if the habitual recovery beverage had been excluded during the training camp. In regards to mitigating markers of muscle damage, a potential positive effect of co-ingestion of protein during exercise may have been visible if we had included elderly subjects who require a greater amount of protein (35–40 g protein) to maximize skeletal muscle protein synthesis rates [47, 48].