Obesity and related cardiometabolic risk factors associated with cardiovascular disease have dramatically increased globally over the recent years [1
]. Obesity is a major public health challenge in Australia and its prevalence for 2019 is estimated to be 67% (32% obese, 34% overweight) [2
]. In Queensland, two in three adults (66%) were overweight or obese in 2017–2018 [2
], signifying a persistent public health problem. To date, no country has successfully reversed this trend [3
]. Population-based preventive measures have been mostly unsuccessful [4
]. Hence, effective treatment programs for those who are already overweight or clinically obese are needed to reduce weight and improve their cardiometabolic profile.
Ketosis is an alternative energy state when glucose levels are low [5
]. Very low-carbohydrate diets, or ketogenic diets are becoming widely recognized as an effective method for treating obesity and managing weight [6
]. Evidence suggests that low-carbohydrate meals lead to greater reductions in body fat and cardiometabolic risk factors, such as fasting glucose levels triglycerides and blood pressure, compared to standard low-fat meals [7
]. This evidence suggests that once tissues become fat adapted utilising ketone bodies as a primary fuel source, profound cardiometabolic benefits emerge.
Despite the cardiometabolic benefits and the greater reduction in body fat, exposure to low-carbohydrate diets has been associated with various adverse outcomes including decreased muscle mass [10
] and reduced bone mass [11
]. Carbohydrate restriction leads to decreases in blood glucose. It is possible that increased gluconeogenic activity could promote the breakdown of muscle tissue and a reduction in muscle glycogen stores [12
]. Reduced muscle mass is an established risk factor for osteoporosis [13
], and has been associated with functional impairment in obese individuals [14
]. Interestingly, the combination of exercise with a low-carbohydrate diet appears to attenuate muscle mass loss induced by ketosis [15
]. However, evidence to date in obese and overweight individuals is limited.
Exercise training has been shown to significantly reduce cardiometabolic risk factors [16
]. In particular the combination of aerobic and resistance training has been associated with increased mitochondrial biogenesis [17
], improved vascular function, lipidaemic profile and reduced inflammation [18
]—all of which are factors known to reduce cardiometabolic disease risk [19
]. However, to date, no study has assessed the combined effect of a low-carbohydrate diet and prescribed exercise on cardiometabolic factors in overweight and obese individuals.
Therefore, the aim of the current study was to assess how an 8 week exercise intervention, consisting of aerobic and resistance exercise, in combination with a low-carbohydrate diet would improve body composition and cardiometabolic risk factors compared to the same exercise intervention combined with the standard Australian dietary guidelines in obese individuals.
A total of 64 obese men and women (BMI, 30.3 ± 3 kg.m−2
; aged, 35.3 ± 9) were recruited from the local community to the ‘Healthy Eating and Living Study’ (HEALS) before being allocated to two conditions by randomised letters: Experimental (structured exercise programme + low-carbohydrate meals; EX-LC) or control (structured exercise programme + standard dietary advice; EX-CO) condition (Figure 1
). Block randomization stratification by gender was undertaken, so that blocks of 12 participants were recruited at a time, randomized into one experimental condition of six participants and one control condition of six participants. Assessors prepared the envelopes with six paper codes (three experimental and three control group) which were added to opaque not concealed envelopes. There were two envelopes: one for females and one for males. Participants were asked to pick one paper from their respective envelope pack and the picked paper would either assign the participant to the experimental or control group. Participants were included if they were aged between 18 and 50 years at enrolment; had a body mass index (BMI) in the range of 30–35 Kg m−2
and no history of bariatric surgery. Participants were excluded if they had been diagnosed of any metabolic or cardiovascular disorder; thyroid disease, autoimmune or neurological disease; had uncontrolled hypertension; were a smoker; were pregnant or breast-feeding or were deemed unfit to exercise by a physician. Participants that were taking drug medication for thyroid, hyperlipidaemia, hypoglycemia, or were taking weigh loss supplements within 3 months before the start of the trial were also excluded. Participants were informed of the methods and study design verbally and in writing before providing written informed consent. The study conformed to the Declaration of Helsinki and was approved by a National Health and Medical Research Council (NHMRC) accredited ethics committee (Bellbery Limited HREC). In addition, the study was registered with the Australian New Zealand Clinical Trials Registry (No.: U1111-1206-4798).
2.2. Study Design Overview
Eligible participants attended the exercise physiology laboratory at the Institute of Health and Biomedical Innovation at Queensland University of Technology at baseline (BL) and 8 weeks after the experimental intervention (POST) each following an overnight fast. Participants were also asked to refrain from alcohol and exercise for 24 h, and caffeine for 12 h, before each visit, as these parameters have been shown to affect primary outcomes of the study [20
]. The laboratory visits included an incremental maximal cycling test for the determination of cardiorespiratory fitness (V
) and a body composition assessment via dual-energy X-ray absorptiometry (DXA), and a venous blood sample was drawn from an antecubital vein in order to assess lipid profile and β-hydroxybutyrate levels. Data collection of all the experimental measurements was performed on the same day for each participant. All participants followed an 8 week (4 sessions per week/45 min per session) prescribed exercise program. The exercise sessions involved a combination of aerobic and resistance exercise that included bouts of moderate [60%–70% HR peak/perceived exertion (RPE) 11–13 on Borg scale] [22
] and high intensity exercise (85%–95% HR peak/RPE 15–17 on Borg scale). During the first 2 weeks of the intervention all sessions were supervised by an exercise physiologist, whilst only one session per week was supervised during weeks 3 to 8. Accredited clinical-exercise physiologists were trained on the prescribed experimental intervention and provided the intervention at the Health Clinics at QUT. The principal investigator of the current trial provided overall supervision of the experimental intervention. Participants were asked to keep a weekly exercise audio diary.
2.3. Dietary Intervention
Participants in the control condition were provided with written information on Australian Dietary guidelines as directed by the NHMRC Australia (NHMRC, 2013). Participants in the experimental condition were provided with 3 pre-prepared meals and 2 interim mid-meal snacks daily during the duration of the trial. A handbook containing recipe examples and food lists based on each experimental condition was also provided to all participants. Neither diet included a specific calorie or energy goal. For implementing controlled feeding protocols for the participants in the EX-LC group, we established a partnership with Thrive Collective Pty Ltd. (Sydney, NSW, Australia), a food service contractor, which provided the pre-prepared meals for the EX-LC group. The low-carbohydrate group pre-prepared meals did not exceed in total 50 g of CHO per day. All foods and quantities consumed were recorded daily for both groups, adherence to the diet was recorded daily via a 24 h food log. Alcoholic beverages were restricted for the intervention period in order to avoid any hyperglycemic incidents in the EX-LC group [23
]. Dietary supplements were not permitted for the 3 months before and during the intervention period. Participants were asked to keep a weekly diet audio diary.
2.4. Maximal Incremental Cycling Test
This test was performed on an electromagnetically braked cycle ergometer (Lode Corival, Groningen, The Netherlands). Following a 3 min warm up (0 W), the test began at 60 W and then increased by 20 W each minute until volitional cessation. Participants self-selected a pedal cadence (>60 rpm) and maintained this throughout the test. Expired gases were collected continuously, and data were averaged every 15 s (Parvo Medics, Salt Lake City, UT, USA) for the determination of oxygen consumption (V
). Peak cardiorespiratory fitness was determined as the highest 15 s average of V
over the last 60 s of maximal exercise (V
peak). Heart rate was measured continuously using a polar heart rate monitor (Polar H10, Kempele, Finland) and recorded, along with perceived exertion (RPE) using the 6–20 Borg scale [22
], during the final 10 s of each stage.
2.5. Body Composition Assessment
Body height was assessed by a standard stadiometer. Body weight was assessed by a digital scale (OMRON, VIVA, Hoffman Estates, IL, USA) before and after intervention. Body fat [total body and regional (VAT; visceral adipose tissue) fat distributions (%)], body mass index (BMI), fat mass index (FMI), lean muscle mass [LMM (cm3)] and total body mineral density [TBMD (g/cm²)] were evaluated following a 12 h overnight fast through DXA (DXA; Hologic QDR 4500 version 12.6) before and after intervention.
2.6. Blood Analysis
Samples were collected in K2 EDTA VacutainerTM tubes using standard aseptic techniques. Plasma was separated by centrifugation (1500× g for 15 min at 22 °C) and stored in 1.5 mL aliquots at −80 °C until further analysis. Fasting blood was analysed before and after intervention for Glucose (BGL), total cholesterol (TC), high density lipoprotein (HDL-C), Triglycerides (TG), C-reactive protein (CRP), low density lipoprotein (LDL-C), Adiponectin and ketone bodies (β-hydroxybutyrate). TC, HDL-C, TG, and CRP results were obtained by spectrophotometric assays on a Cobas Integra 400 (Roche, Rotkreuz ZG, Switzerland); LDL-C was then calculated by the Friedewald formula using the measured TG, TC, and HDL-C results. Adiponectin was measured by enzyme-linked immunosorbent assay (ELISA) (Millipore, Billerica, MA, USA) and ketone bodies (β-hydroxybutyrate, βHB) were obtained on a Beckman Coulter AU400 analyser (Beckman Coulter, Brea, CA, USA) utilising a RANBUT Test Kit (Randox, Crumlin, UK).
2.7. Resting Blood Pressure
Brachial systolic and diastolic blood pressure was measured using a manual analogue sphygmomanometer (Omron, Exactus Aneroid Sphygmomanometer, Melbourne, Australia) before and after intervention after ten minutes of seated rest.
3. Statistical Analysis
Based on the reported effect size (ES = 7.9 ± 1.6) on a previous metanalysis that assessed percent change of body fat in diet and exercise conditions [24
], the current study adopted a conservative effect size of d = 0.80, with alpha at 0.05, and power at 0.90 which resulted in a sample size of 30 participants for each condition. The data were normally distributed. Our assessment of kurtosis and skewness of the participants’ age and body composition revealed values between −2 and 2 and therefore our data distribution is considered normal [25
]. A single-factor linear mixed model was used to compare anthropometric characteristics, blood biomarkers and cardiorespiratory fitness between the experimental and control conditions. A two-factor (condition*time) linear mixed model was used to detect differences in body composition, blood biomarkers and cardiorespiratory fitness across time (before and after trial) and between conditions. Post-trial data were also analysed as changes from baseline (delta) to account for individual baseline variance. Statistically significant interactions were further investigated with multiple comparisons using Fisher’s least significant difference approach [26
]. Our cohort was further stratified based on the β-hydroxybutyrate (βHB) ketone levels (βHB of 0.3 mmol/L≥ achieved ketosis; BHB of 0.3≤ not in ketosis) [5
] and a multiple regression analysis was performed to investigate the mechanisms influencing the observed post-intervention change in the examined indices. Analyses were conducted using the Statistical Package for Social Sciences (Version 22; IBM SPSS Inc., Chicago, IL, USA) and statistical significance was set at p
≤ 0.05. Data are presented in the text and Tables as the mean and standard deviation (SD) unless otherwise stated.