Physical activity and the type of diet are important factors regulating the metabolism of the body and thus affecting the state of human health. In the search for a type of diet that could improve human health, some advantages of low-carbohydrate diets (LCDs) and especially its variety called the Atkins diet were noted. In Poland, the Atkins diet was modified by Dr Kwaśniewski [1
], who called it an “optimal diet”. This diet maintains the proportion of proteins/fat/carbohydrate in the range of 1:2.5–3.5:0.5, whereas it does not restrict total energy intake. Moreover, this diet eliminates the following foods: honey, jam, sweets, sucrose, bread, white rice, beans, starch, potatoes and sweetened drinks.
LCDs are similar to high-fat diets (HFDs) and are a broad category lacking an objective definition. The carbohydrate amount of 45–65% of the total daily energy intake is suggested as appropriate for adults [2
]. Diets with intakes below 45% can be viewed as LCDs. These diets have also been defined as having an upper limit of 40% of total daily energy from carbohydrate [3
] or having less than 200 g of this dietary ingredient (3). A more restricted term for LCD suggests specifying non-ketogenic LCD, which contains 50–150 g of CHO, because pure ketogenic diets have a maximum of ~50 g or ~10% of total daily energy from this ingredient [5
]. This diet maintains protein at a moderate level (1.2–1.5 g/kg/day) with predominance of energy intake from fat (~60 to 80% or more) [6
]. In the light of these criteria, the Atkins and Kwaśniewski diets can be partially included in the ketogenic diets, which ought to be confirmed by a significant increase in the concentration of β-hydroxybutyrate in blood [7
Most studies connected with LCDs have been performed on sick individuals, because this diet is believed to improve the lipid profile and glycemic control [9
]. Moreover, LCDs with energy restriction are used by obese or overweight subjects in weight loss programs [12
]. Soenen et al. [14
] demonstrated that the higher protein content of LCDs rather than lower carbohydrate amount in this diet was the crucial factor in greater weight loss during applied hypocaloric nutrition. It is worth pointing out that non-calorically restricted ketogenic diets have also led to body fat and/or body weight reduction [15
]. However, this effect seems to occur as a spontaneous energy intake reduction, which could increase satiety through suppression of ghrelin production [17
]. Furthermore, it was found that a ketogenic diet does not increase cardiac risk factors in hypercholesterolemic people [18
], and an LCD with energy intake coming from CHO below 20% also showed no negative cardiovascular risk in obese individuals with type 2 diabetes [19
]. Another field of interest concerning a beneficial impact of a ketogenic diet is its influence on athletes’ exercise performance [20
]. However, in contrast to the proposed benefits of fat adaptation for physical performance, another study showed increased fat oxidation but high-intensity work output was impaired [15
High fat intake leads the body to become fat-adapted or keto-adapted. However, so far, long- term studies lasting only up to 2 years have been performed in individuals affected by so-called non-communicable diseases. The main objective of this observational retrospective study was to determine whether 3 years’ adherence to an LCD is associated with metabolic, cardiovascular, plasma lipids and somatic variables of middle-aged men who began to consume this diet as healthy subjects, being convinced of its beneficial protection against susceptibility to disease. Their counterparts were volunteers with matched age, weight and height who used mixed diets (MDs). The intervention part of this study consisted of graded exercise performed to individual maximum load. Our assessment was focused on a survey of risk factors of metabolic and cardiovascular diseases, plasma lipids as well as somatic and exercise capacity variables.
2. Material and methods
Fifteen apparently healthy men who self-reported adherence to LCD for at least 3 years volunteered to participate in this study. All participants had current medical examinations, without any contraindications to performing exhaustive exercise. After medical examination performed by a general practitioner 3 subjects were eliminated from the study, so 12 participants took part in this observation. They declared that they had never engaged in regular physical activity of moderate to vigorous intensity. However, we did not control the varied levels of fitness among the participants. The LCDs subjects were members of local supporters belonging to a nongovernmental society called the “All-Polish National Association of Optimal Brotherhoods”. They declared that they had maintained an LCD for at least 3 years (mean = 4.58 ± 1.1, min = 3, max = 6.5 years). The control group for subjects who applied an LCD comprised 12 volunteers with matched age, weight and height who used an MD all the time. All of the study participants were informed of the objective of the experiment and the accompanying risks. Volunteers provided their written, voluntary, informed consent before participation.
Inclusion criteria were: (1) LCD for at least 3 years; (2) age 40–60 years; (3) BMI 20–29.9 kg/m2; (4) body mass 50–90 kg; (5) the lack of chronic diseases; (6) systolic blood pressure 100–140 mmHg and diastolic blood pressure 60–90 mmHg. The study excluded participants with (1) using drugs, drinking alcohol and smoking; (2) hypertension; (3) prematurely stopped exercise test.
The research project was conducted according to the Helsinki Declaration and was approved by the Ethics Committee for Scientific Research at the Jerzy Kukuczka Academy of Physical Education in Katowice, Poland.
2.2. Experimental Design
All healthy participants came to the laboratory in the morning, between 8:00 and 9:30 AM, after an overnight fast and abstention from alcohol, medications and exercise for 2 days. In the first stage of the study age and basal somatic data (body height—BH, body mass—BM, body fat—BF, free fat mass—FFM, total body water—TBW and body mass index—BMI) were recorded. Variables were estimated by bioelectrical impedance analysis using the Tanita Body Fat Analyzer TBF 300A (Tanita, Amsterdam, Netherlands). All participants provided to the laboratory a 7-day dietary enrollment recorded with a 24-h dietary recall form completed to assess their habitual daily energy and nutrient intakes. All nutrient data were analyzed using the National Food and Nutrition Institute computer database (Dietus, BUI INFIT, Warsaw, Poland).
Before the incremental exercise test the rest heart rate (HR) and blood pressure (BP) were measured (Oxycon-Alpha ER 900, Jaeger, Hoechberg, Germany). Also, blood samples from the antecubital vein were drawn for determination of concentrations of the following biochemical variables: glucose (G), lactate (LA), uric acid (UA), β-hydroxybutyrate (β-HB), free fatty acids (FFA), total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), triacylglycerols (TG), and immunoreactive insulin (IRI). Low-density lipoprotein cholesterol (LDL-C) concentration was calculated.
Then after a 10-min rest the blood pressure (systolic—SBP and diastolic—DBP), HR, oxygen uptake (VO2), carbon dioxide excretion (VCO2) and respiratory exchange ratio (RER) were recorded. On the basis of the obtained data, the following physiological variables were calculated: mean arterial pressure (MAP), pulse pressure (PP), and rate pulse pressure product (RPP). After this time, the subjects from both groups sat on the cycloergometer (ER 900, Jaeger, Hoechberg, Germany) and started work at a rate of 60 revolutions per minute, beginning from 0 W. The load was increased by 30 W to an individual maximum fatigue, maintaining each work stage for 3 min. At maximal load, the above-mentioned physiological variables were recorded and calculated. At each submaximal load HR, VO2, VCO2 and RER were also recorded. Expired air was analyzed at rest and during an incremental test using the cycloergometer and an Oxycon-Alpha quick gas analyzer (Jaeger, Hoechberg, Germany).
Biochemical analyses were determined in fasting blood samples, which were collected in heparin- or EDTA-treated tubes. Serum and blood plasma were separated, and were assayed immediately for determination of glucose (G), uric acid (UA) and LA concentrations using diagnostic kits: GL 2623 and UA 230 from Randox and BioMérieux Laboratories Ltd., respectively (Spectrophotometer UV-VIS 1202, Shimadzu. Blood samples for determination of β-HB were rapidly deproteinized by addition of 0.6 N perchloric acid. Protein free supernatants, part of plasma and serum were stored at −80 °C and analyzed using a commercial (RANBUT) Randox kit. Plasma TC, HDL-C, TG and serum FFA were determined with enzymatic methods using commercial Randox kids (CH 200, CH 203, TR 1697, and FA 115, respectively) (Spectrophotometer UV-VIS 1202, Kyoto, Shimadzu). Level of low-density lipoprotein cholesterol (LDL) was calculated using the Friedewald formula [25
]. Risk for cardiovascular disease (CVD) was evaluated by calculation of the ratios of TC/HDL-C (R1
), LDL-C/HDL-C (R2
), and TG/HDL-C (R3
Serum IRI concentration was quantified by the electrochemiluminescence method using an Elecsys 1010 analyzer (Roche Diagnostics, Mannheim, Germany). From the ratio of multiplication fasting G (millimole per liter) and IRI concentrations (milliunits per liter) divided by 22.5 we calculated the homeostasis model assessment (HOMAIR
]. Both indices, i.e., TG/HDL-C and HOMAIR
, were applied as surrogate measures of insulin resistance [26
2.3. Statistical Analyses
Results were reported as mean values ± SD. The Shapiro-Wilk test was performed to verify the distribution of the variables. The nonparametric Mann-Whitney U test and two-way ANOVA test (for repeated measures) with post hoc Bonferroni test were used to assess the differences between variables of both investigated groups. The tests were set with a confidence interval of 95% and differences with p < 0.05 were accepted as statistically significant. The statistical analyses were conducted with STATISTICA 12.0 software version (StatSoft, Krakow, Poland).
Subjects who participated in the study reported adherence to the high-fat LCD for more than 3 years (mean = 4.58 ± 1.1, min = 3, max = 6.5 years). Basic somatic characteristics for the two study groups are presented in Table 1
. Age of both groups was similar and none of the somatic data differed significantly.
It was found that average energy intake was limited to approximately 2075 kcal/day in the LCDs subjects and 1870 kcal/day in the MDs group, and did not differ significantly (Table 2
Average CHO intake was limited approximately to 118 g/day (23% total daily energy intake) in the LCDs group and was lower than in the MDs subjects—229 g/day (49% total daily energy intake)—p
< 0.001. Moreover, the average fat intake was approximately 150 g/day (65% total daily energy intake) in LCDs participants and 77 g/day (37% total daily energy intake) in MDs men, whereas protein intake was a little lower than the recommended dietary allowance (it ranged from approximately 0.89 g/kg bm
to 0.93 g/kg bm
and did not differ significantly between the groups) [27
Mean values for FFA, β-HB, G, UA, IRI and LA were measured in blood samples and are given in Table 3
. FFA (in all samples), β-HB (in LCDs men), G (in LCDs group at maximal exercise) markedly exceeded upper borderline levels. Kind of applied diet (F = 3.45, p
< 0.05) and maximal physical effort (F = 3.72, p
< 0.05) were associated with increased plasma FFA concentration and post hoc analysis showed that this level in the LCDs group at maximal exercise was significantly higher than in the MDs group (Cohen’s d = 1.52) and at rest (Cohen’s d = 0.44) in the first abovementioned group (p
< 0.01). LCDs were associated with increased β-HB plasma concentration (F = 20.4, p
< 0.001) and it was significantly higher in the LCDs group than in the MDs group at rest (Cohen’s d = 3.2) and maximal exercise (p
< 0.01; Cohen’s d = 0.32). Moreover increased blood LA concentration was associated with applied physical effort (F = 104.60, p
< 0.001). LA concentration in blood at maximal exercise level was significantly higher than at rest in both investigated groups (p
< 0.001; LCDs—Cohen’s d = 2.98; MDs—Cohen’s d = 3.8).
presents the levels for principal biomarkers of CVD risk associated with traditional lipid parameters (TC, HDL-C, LDL-C, TG, TC/HDL-C, LDL-C/HDL-C, TG/HDL-C), HOMAIR
, at rest and maximal intensity exercise. TC levels in LCDs subjects at rest and maximal exercise, and in MDs men under a maximal effort; HDL-C concentration in LCD group at maximal exercise; LDL-C serum concentration in LCDs men at rest and maximal exercise and in the MDs group at rest, and TC/HDL-C ratio calculated at rest in the LCDs group exceeded the upper reference limits.
The kind of applied diet was associated with increased TC concentration (F = 21.8, p < 0.001) and it was higher at rest (Cohen’s d = 1.84) and maximal exercise (Cohen’s d = 1.73) in LCDs subjects than in MDs participants. Physical effort was also associated with increased serum concentration of this variable (F = 17.87, p < 0.001), where this level in LCDs subjects at maximal exercise was higher than at rest (p < 0.05; Cohen’s d = 0.82). Analysis of variance indicates the association between diet (F = 8.95, p < 0.01) and physical exercise (F = 16.99, p < 0.001) with the level of HDL-C. It was higher (p < 0.05) at rest in LCDs subjects than in the MDs group (Cohen’s d = 0.48), and exercise level in LCDs subjects was significantly higher in comparison with resting values (p < 0.01; Cohen’s d=0.4). Moreover, used diets were associated with increased LDL-C serum concentration (F = 20.4, p < 0.001), and at maximal exercise this level was higher in LCDs men than in MDs subjects (p < 0.01; Cohen’s d = 1.74).
Maximal workload during the incremental test (WRmax) was significantly lower in the LCDs group (145.00 ± 28.1 W) than in the MDs group (175.00 ± 35.8 W), and total work (TW) reached during the ergocycle test (reached the levels-USUNĄC) was 78.3 ± 29.4 kJ in the LCDs group and 112.2 ± 42.2 kJ in MDs subjects, which were significantly different (p < 0.05).
Participants had a normal resting reference range of HR (<90 bpm), SBP (<130 mm Hg), DBP (<85 mm Hg, except MD group), MAP (<105 mmHg), PP (<63 mmHg) and RPP (<15 bpm × mmHg), whereas these parameters reached individually different levels during maximal effort. Two-way analysis of variance with repeated measures indicated a significant influence of physical effort on HR (F = 498.8, p
< 0.001), SBP (F = 142.01, p
< 0.001), MAP (F = 46.16, p
< 0.001), PP (F = 148.16, p
< 0.001), and RPP (F = 412.71, p
< 0.001). Post hoc analysis also showed (Table 5
) that these parameters measured or calculated at rest were significantly lower than at maximal physical effort in both groups (p
Mean values of HR and respiratory variables presented in Table 6
recorded at rest and during maximal exercise were within the reference range. The analysis of variance revealed the association between applied diet on HR (F = 9.6, p
< 0.01), VO2
(F = 13.57, p
< 0.01) and RER (F = 15.75, p
< 0.001). Post hoc calculations showed that between-group differences were present at rest, 30 W, 60 W, 90 W, 120 W and maximal load, in relation to HR, VO2
and RER, respectively. Moreover, significant changes in the above-mentioned variables were observed under the influence of physical effort in HR (F = 216.1, p
< 0.001), VO2
(F = 781.18, p
< 0.001) and RER (F = 69.77, p
< 0.01). In post hoc analysis there were observed significantly higher exercise values of HR, VO2
and RER in comparison to their resting level in both study groups.