Carbohydrates and exercise performance in non-fasted athletes: A systematic review of studies mimicking real-life
Abstract
There is a consensus claiming an ergogenic effect of carbohydrates ingested in the proximity of or during a performance bout. However, in performance studies, the protocols that are used are often highly standardized (e.g. fasted subjects, constant exercise intensity with time-to-exhaustion tests), and do not necessarily reflect competitive real-life situations. Therefore, we aimed at systematically summarizing all studies with a setting mimicking the situation of a real-life competition (e.g., subjects exercising in the postprandial state and with time-trial-like performance tests such as fixed distance or fixed time tests). We performed a PubMed search by using a selection of search terms covering inclusion criteria for sport, athletes, carbohydrates, and fluids, and exclusion criteria for diseases and animals. This search yielded 16,658 articles and the abstract of 16,508 articles contained sufficient information to identify the study as non-eligible for this review. The screening of the full text of the remaining 150 articles yielded 17 articles that were included in this review. These articles described 22 carbohydrate interventions covering test durations from 26 to 241 min (mostly cycling). We observed no performance improvement with half of the carbohydrate interventions, while the other half of the interventions had significant improvement between 1% and 13% (improvement with one of five interventions lasting up to 68 min and with 10 of 17 interventions lasting between 70 and 241 min). Thus, when considering only studies with a setting mimicking real-life competition, there is a mixed general picture about the ergogenic effect of carbohydrates ingested in the proximity of or during a performance bout with an unlikely effect with bouts up to perhaps 70 min and a possible but not compelling ergogenic effect with performance durations longer than about 70 min.
Keywords: Carbohydrates; Performance; Sport; Athlete; Exercise; Ergogenic; Time trial; Athletic performance; Dietary supplement
Introduction
Carbohydrates are one of the two main fuels for sport activities, and their relevance for optimal sport performance is undisputed among experts [1]. Athletes not only ingest carbohydrates as general contributors to their daily energy need but also specifically as ergogenic agents in a more time-specific way, such as during a sport event or in the days preceding it. This potential ergogenic effect of carbohydrates has been the subject of numerous investigations, and a series of reviews have summarized their outcome (e.g., [1,2]). In general, there is a consensus claiming an ergogenic effect of carbohydrates ingested just before or during a performance bout.
One of the fundaments of science is the continuing questioning of current theories in order to corroborate them or, in the case of contradicting new evidence, to challenge them. In the case of the ergogenic effect of carbohydrates ingested near to a performance bout, one could question if the study designs used to investigate this effect were suitable for extrapolating their outcomes to a real-life situation, particularly as a standardized, controlled laboratory setting can be quite different from the conditions of a real-life situation.
Subjects often fast overnight in performance studies. The reason for this is probably that the metabolism in fasted subjects is in a more balanced state, which might be more easily reproduced than a postprandial state. However, the recommendation to athletes is not to compete in a fasted state because of potentially reduced liver glycogen stores and a subsequent negative effect on performance [3]. While this concern might be unsubstantiated (overnight-fasted well-trained subjects can have more than twofold higher liver glycogen levels compared to overnight-fasted non-athletes, ca. 130 g [4] vs. ca. 50 g [5], respectively), athletes almost intuitively do not compete in a fasted state. Further, a test mode assessing how long a subject can exercise at a given intensity is common in “performance” studies (e.g., time-to-exhaustion tests). This is also does not reflect the real-life situation as usually a sporting event, at least in elite sports, requires performing either as fast as possible for a given distance (e.g., races) or as well as possible within a given time (e.g., team sports). See Currell and Jeukendrup for a discussion on the different types of performance tests [6].
In two recent meta-analyses, the ergogenic effect of carbohydrates ingested during endurance sport activities was investigated with the inclusion of fasted subjects [7,8]. In our study, we also aimed at systematically reviewing the influence of carbohydrate intake on performance, but as we wanted to focus on the real-life applicability of the study outcome, we defined a priori to exclude studies with subjects who were fasted and in which the performance test was of a time-to-exhaustion character. However, it was not the focus of this review to discuss why carbohydrate ingestion in the proximity of a performance bout may or may not help athletes.
Methods
We defined the protocol for conducting the systematic search before commencing the data search and we did not modify it thereafter. One of us (PCC) then screened the PubMed for studies with a carbohydrate intake during the days prior to a sport performance, i.e., a carboloading regime, and studies with a carbohydrate supply immediately before or during a performance bout.
We originally planned to conduct a series of meta-analyses but abandoned this idea when it became clear that the number of studies fulfilling our inclusion criteria was small and the study designs were too heterogeneous for a meaningful grouping.
Data sources and search strategy
The PubMed was searched up to September 3, 2011 with the following keyword combination: (Exercise OR Sport OR Athlete OR Athletes) AND (Hydration OR Water OR Fluid OR Drink OR Drinks OR Beverage OR Beverages OR Glycogen OR Loading OR Carbo OR Carbohydrate OR Carbohydrates OR Glucose OR Fructose OR Maltodextrin) NOT (Mice OR Mouse OR Pig OR Pigs OR Rat OR Rats OR Horse OR Horses OR Fish OR Dog OR Dogs OR Patient OR Patients OR Disease OR Diseases OR Diabetes OR Obesity OR Obese OR “Cord injury” OR Wheelchair). The plural of several terms was included as we realized that using only the singular of that term yielded a different number of hits (although using the singular of a term should actually also identify abstracts including only the plural of a term). We discarded articles with sufficient information in the abstract that clearly identified the study as ineligible, without consulting the full-text of the article. For the remaining articles, we checked the full-text for the information needed to evaluate the study. We additionally consulted the list of articles judged as eligible for the two meta-analyses mentioned above [7,8].
Inclusion criteria
We included studies with a randomized, crossover, placebo-controlled, and if possible blinded study design. Blinding was not feasible as an absolute criterion, as sometimes the intervention could not be fully masked (e.g., carboloading vs. no carboloading). Additional criteria were a mean age of the subjects between 18 and 40 years (both genders allowed), a reported VO2max ≥ 50 mL/kg/min body mass, and a performance test in the postprandial state (2 to 4 h after ingesting last meal). The performance test had to be either of a time trial (TT) character, i.e., a fixed distance, fixed time, or fixed amount of work, or a submaximal exercise followed by a TT (S+TT). Further, for studies with a carbohydrate intake immediately prior to and/or during the exercise, we included only studies with provision of any type of carbohydrates, electrolytes, and water, but no further components (e.g., caffeine, choline, protein, amino acids, and fatty acids). We excluded studies with time-to-exhaustion tests assessing the exercise capacity and discarded all studies without sufficient methodological information to enable a check of the inclusion criteria.
Grouping of interventions
We combined the interventions according to sport type and physiological impact of the activity (in particular with respect to endogenous energy delivery). The three grouping factors were the test mode (cycling, running, or other), a similar test duration (up to 60 min, 61 to 90 min, 91 to 120 min, 121+ min), and a comparable carbohydrate intervention (carboloading vs. no carboloading, ingestion of carbohydrate containing drinks vs. drinks containing no carbohydrates).
Performance
The outcome was the overall performance during the TT or the TT part in the S+TT interventions as time needed to cover the fixed distance, distance covered within a fixed time, or power accomplished within a fixed time.
Results
Eligible studies and interventions
The PubMed search yielded 16,658 articles and the abstract of 16,508 articles contained sufficient information to identify the corresponding study as not eligible. The full text of the remaining 150 articles allowed the identification of 16 articles that fulfilled the inclusion criteria. We identified one additional eligible article [9] through a scan of the reference lists of both meta-analyses of the similar topic [7,8]. Overall, the 17 articles comprised 22 interventions (14 articles with one intervention, one article with two interventions, and two articles with three interventions, Figure 1).
thumbnailFigure 1. Intervention type, performance bout and outcome of eligible studies.
Type of studies
Two studies were carboloading interventions; in one study the drink was not ingested (mouth-rinse study), and in the remainder 19 interventions, the effect of a carbohydrate-containing drink was compared to a non-carbohydrate placebo. In four of these 19 interventions, the carbohydrate type was n