The Effects of Carbohydrate versus Fat Restriction on Lipid Profiles in Highly Trained, Recreational Distance Runners: A Randomized, Cross-Over Trial

Abstract

A growing number of endurance athletes have considered switching from a traditional high-carbohydrate/low-fat (HCLF) to a low-carbohydrate/high-fat (LCHF) eating pattern for health and performance reasons. However, few studies have examined how LCHF diets affect blood lipid profiles in highly-trained runners. In a randomized and counterbalanced, cross-over design, athletes (n = 7 men; VO2max: 61.9 ± 6.1 mL/kg/min) completed six weeks of two, ad libitum, LCHF (6/69/25% en carbohydrate/fat/protein) and HCLF (57/28/15% en carbohydrate/fat/protein) diets, separated by a two-week washout. Plasma was collected on days 4, 14, 28, and 42 during each condition and analyzed for: triglycerides (TG), LDL-C, HDL-C, total cholesterol (TC), VLDL, fasting glucose, and glycated hemoglobin (HbA1c). Capillary blood beta-hydroxybutyrate (BHB) was monitored during LCHF as a measure of ketosis. LCHF lowered plasma TG, VLDL, and TG/HDL-C (all p < 0.01). LCHF increased plasma TC, LDL-C, HDL-C, and TC/HDL-C (all p < 0.05). Plasma glucose and HbA1c were unaffected. Capillary BHB was modestly elevated throughout the LCHF condition (0.5 ± 0.05 mmol/L). Healthy, well-trained, normocholesterolemic runners consuming a LCHF diet demonstrated elevated circulating LDL-C and HDL-C concentrations, while concomitantly decreasing TG, VLDL, and TG/HDL-C ratio. The underlying mechanisms and implications of these adaptive responses in cholesterol should be explored.

Keywords: cholesterol; crossover; endurance athletes; lipid profiles; low-carbohydrate; low-fat.

Figure 1

Experimental Approach. In a randomized, counterbalanced, cross-over manner, seven recreationally trained endurance athletes were assigned to consume either a low-carbohydrate/high-fat diet (LCHF) or high-carbohydrate/low-fat diet (HCLF) for six weeks. A two-week washout was allowed between each dietary phase. All participants maintained a constant exercise regimen and a 3-day food log throughout the experimental phases. Repeat intravenous and capillary blood measures were collected on days 4, 14, 28, and 42 during each feeding condition. The exploratory goal of this study was to assess the main effects and interactions of LCHF and HCLF on comprehensive blood lipid and glucose panel markers. This figure was created with BioRender^®.

Figure 2

Fasted Capillary Ketone Responses. Each data point denotes a 7-day BHB average. The LCHF feeding phase modestly elevated circulating fasted BHB into the range of nutritional ketosis. Average weekly BHB plateaued at 0.5 ± 0.05 mmol/L (range: 0.20–1.24 mmol/L) over the six weeks. A 1 (condition) × 6 (time) RM ANOVA revealed that BHB concentrations remained stable over time from WK1 and thereafter (p = 0.29).

Figure 3

Mean participant energy intake and macronutrient distribution. The 3-day dietary records collected from all participants (n = 7) are presented as a weekly average of percent energy distribution from dietary fat, protein, and carbohydrates. A one-way ANOVA between diet conditions revealed significant differences across all relative macronutrient categories (p < 0.001), but no significant differences in total energy intake (p > 0.05).

Figure 4

Baseline normalized change in plasma lipid markers. (A) total cholesterol, (B) HDL-C cholesterol, (C) LDL-C cholesterol, (D) triglycerides, (E) triglyceride to HDL-C ratio, and (F) total cholesterol to HDL-C cholesterol ratio. ANOVA revealed a main effect of diet in all the markers analyzed († = p < 0.05). Overall, LCHF diet lowered circulating triglycerides and TG/HDL-C ratio while simultaneously increasing HDL-C when compared to the HCLF diet. A significant interaction was detected for total cholesterol, LDL-C, and TC/HDL-C ratio. Compared to HCLF, the LCHF diet elicited significantly higher cholesterol and LDL-C values, and greater TC/HDL-C ratios at day 14 and thereafter. Main effects: Time: *, *** p < 0.05/0.001 compared to Day 4; Condition effects: † = p < 0.05; Interaction: # = p < 0.05 between conditions at indicated timepoints. Data presented as mean ± SD; n = 7.