Experimental overview
This study used a randomized, crossover design with participants reporting to the laboratory on three days (Fig. 1). The first visit consisted of an incremental swimming test to volitional fatigue in a swim flume to determine maximal aerobic capacity (VO2max). The second and third visits consisted of submaximal swimming at velocities equivalent to 50, 60, and 70% of VO2max as determined during the first visit. Prior to the second and third visits, participants consumed either a HCLF or a LCHF diet for three days. A non-ketogenic LCHF diet was chosen a) to study its use as the type of short-term dietary change that would be included in a periodized nutrition plan, and b) to remove the confounding effect of ketones on the relationship between respiratory exchange ratio and substrate utilization during exercise (19). Dietary interventions were randomly assigned and separated by a 4-day washout period.
Participants
Eight healthy, recreational swimmers (4 male, 4 female, 34.6 ± 9.4 years old, BMI 23.8 ± 2.6 kg m-2, VO2max 42.4 ± 8.5 ml kg-1 min-1, average swim training volume 5.2 ± 2.3 h per week) were recruited from a local swim club. To be included in the study, participants had to be between the ages of 18-59 years old, swimming >3 km per week, and be willing to manipulate their dietary patterns for two weeks. Participants were informed of the risks and benefits of participating and provided written informed consent before participating in the study. This study was approved by the California State University, Los Angeles Institutional Review Board (protocol #1419221).
Procedures
Height and weight were collected on their first visit using a high-capacity column scale (Seca 703, Hamburg, Germany). Participants were instructed to refrain from vigorous exercise for 24 hours prior to each visit. All swim tests were completed using a 4.27 x 2.13 m outdoor swimming flume (Endless Pools, Aston, PA, USA). The water depth was maintained at 1.14 m. Water temperature was maintained at 27 °C.
During the first visit, subjects completed an incremental swimming test to exhaustion for determination of VO2max. Following a self-selected 5-10 min warm-up in a 22.9-m outdoor pool, participants transitioned to the flume for a 1-2 min familiarization swim. For the graded exercise test, the initial intensity was set at 0.93 m s-1 (1:47 per 100 m) with the speed increasing by 0.09 ± 0.01 m s-1 every 2 min. The test was terminated when participants could no longer maintain pace or when they reached volitional fatigue. Heart rate (HR) was continuously monitored via telemetry (Polar T31, Kempele, Finland). Breath-by-breath gas exchange data were continuously measured using a metabolic cart (Quark CPET; Cosmed, Rome, Italy). The VO2max was determined as the highest 10-s average. The gas analyzers on the metabolic cart were calibrated to ambient air and certified standard gas of known concentration (16% O2, 5% CO2, 79% N2) and gas volume was calibrated with a 3-L syringe.
For the second and third visits participants arrived in a fasted state, with trials performed in the morning at the same time of day following both diet interventions. Participants were allowed a self-selected 5-10 min warm-up in the pool prior to the start of the test. Swimming speeds were established based on the speeds eliciting 50, 60 and 70% of their VO2max during the incremental test on the first visit. The subjects swam at each speed for five minutes and each trial was separated by a brief rest period to drain the snorkel of any collected fluids and provide water ad libitum. Heart rate and breath-by-breath respiratory measurements were continuously collected throughout the exercise trial. The rates of O2 consumption (VO2) and CO2 production (VCO2) were determined from the average values over the last two minutes of each 5-min stage. Rate of energy expenditure (EE) was calculated according to the equation of Péronnet & Massicotte (20) and assumed negligible protein oxidation:
The rate of EE was divided by the swimming speed to determine Cs. Oxygen pulse, a non-invasive estimate of stroke volume was calculated for each participant by dividing VO2 by heart rate (21).
Dietary intervention
Prior to enrollment participants were asked to complete a 3-day dietary food recall on three consecutive weekdays (Wednesday – Friday). Participants were provided with diet record sheets and instructed to accurately record all food and drinks consumed with estimates based on basic household portion sizes. Based on their individual eating behaviors and dietary preferences, individualized HCLF (70% carbohydrate, 15% fat, 15% protein) and LCHF (15% carbohydrate, 70% fat, 15% protein) diets were created using dietary analysis software (ESHA Food Processor Nutrition Analysis, Salem, OR, USA) and provided to each participant. Total energy intake was determined as the average of the Harris-Benedict (22) and Mifflin-St. Jeor (23) resting energy expenditure equations, multiplied by an activity factor of 1.55. Participants received individualized counseling and instruction on how to follow the diets, as well as basic kitchen measuring equipment and a food-safe digital scale (Etekcity, Digital Kitchen Scale, EK6015, Anaheim, CA, USA). Diet instructions were provided on a Monday and participants were told to follow them as closely as possible for the subsequent Wednesday–Friday, before arriving in the fasted-state on Saturday morning for testing. Subjects noted any deviations from their prescribed diets. Seven participants followed the standard protocol with a 4-day washout period between dietary interventions, but due to scheduling conflict one participant had an 11-day washout period.
Statistical analysis
All data are reported as means and standard deviations. Dietary intake is reported as the average over three days and analyzed with a paired t-test, after confirming normality of the data. Repeated measures analysis of variance (2 diets x 3 exercise intensities) determined main effects for diet and exercise intensity on physiological endpoints with Bonferroni corrections used for post hoc testing. The Greenhouse-Geisser correction was used when sphericity was not met. Analyses were performed using Jamovi (Version 1.2.16.0, www.jamovi.org) with statistical significance at p < 0.05.