Global obesity rates continue to rise, especially in developing Asian countries and among women (1), which consequently increases the risk of developing metabolic diseases such as diabetes and cardiovascular disease (2). Excessive fat accumulation in those with obesity is considered a major cause of insulin resistance characterized by increased postprandial glucose and insulin responses, and consequent metabolic endotoxemia, irregular androgenicity, and decreased cardiometabolic capacity (3–6). Fortunately, lifestyle, exercise, and nutrition interventions reverse such deleterious outcomes, especially in high-risk individuals, including women with obesity and associated metabolic abnormalities (7–9).
High-intensity interval training (HIIT) has recently been proposed as a weight and fat loss strategy by modulating insulin sensitivity and associated cardiometabolic biomarkers (8,10,11). In young healthy women, 15 wk of HIIT reduced body fat percentage (BF%) and enhanced insulin sensitivity more than a similar frequency lower-intensity exercise (8). However, in women with overweight, no improvement in insulin sensitivity was found despite reduced total and abdominal BF% (11). Concerns over HIIT safety, long-term benefits, adherence, and adverse physiological side effects in high-risk individuals and those with obesity continue to be debated (12). Women with obesity are also exposed to obesity-related abnormal androgenicity, especially irregular testosterone levels, which are linked to insulin resistance and cardiovascular risks (13). Limited recent evidence from animal models showed that HIIT could not improve obesity-induced inhibitory effects on sex hormone development (e.g., kisspeptin transcription and gonadotropin-releasing hormone) (14). Whether HIIT ameliorates such interrelated obesity-induced metabolic and hormonal irregularities is unknown in humans, especially in women. Furthermore, the development of insulin resistance is often concurrent with associated cardiometabolic abnormalities such as bacterial metabolic endotoxemia, which is common in those with obesity, diabetes, and cardiovascular diseases (15). Evidence in rodents and humans suggests a role for gut-derived endotoxins (e.g., lipopolysaccharide) in obesity-related low-grade chronic inflammation (15). Whether and how HIIT affect endotoxins in individuals with obesity is currently unknown.
Caffeine is known to facilitate fat loss when combined with exercise components (16–18). Both chronic and acute intakes of caffeine have been implicated in lowering glucose, enhancing insulin sensitivity, and preventing type 2 diabetes (19–21). Whether HIIT combined with caffeine intake improves cardiometabolic and reverses androgenicity outcomes in women with obesity has never been answered.
The present study examines the effectiveness of HIIT intervention with either caffeine (CAF) or placebo (PLA) in Taiwanese women with obesity. The primary outcomes tested are body composition, and the secondary outcomes are glucose and insulin responses under oral glucose challenged condition, endotoxins, testosterone, and cardiorespiratory and anaerobic fitness. We hypothesized that disparate HIIT-induced cardiometabolic, hormonal, insulin resistance, and endotoxic responses are optimized when CAF is supplemented with HIIT in women with obesity.
MATERIALS AND METHODS
Design and Participants
The study is a randomized placebo-controlled single-blind intervention trial. Participants were sedentary Asian female volunteers with obesity (age 18–30 yr). Obesity was defined based on the Asian cutoff point of body mass index (BMI) ≥27 kg·m−2 and/or BF% >30% (Ministry of Health, Taiwan, ROC), and participants were recruited using convenient sampling. Exclusion criteria included medication, chronic cardiovascular or metabolic diseases history, or any allergy to caffeine intake. None of the participants reported taking contraceptive pill or exceeded a maximum of two cups of coffee daily. Participants were asked to maintain similar lifestyle habits for diet and physical activity throughout the intervention.
The sample size was calculated based on two separate randomized groups (CAF and PLA) to provide the least meaningful difference on endotoxins (22), to be induced by HIIT, for a power of 90% at 5% significance α level. The required sample was 20 (10 in each group); we recruited 28 to allow for a dropout. Four participants dropped out (three because of time commitment, one after initial assessment), with the remaining number of participants required for the intervention (compliance rate of 86%). Twenty-four participants completed the intervention after being equally randomized into two separate HIIT groups with either PLA or CAF (12 in each group). All participants performed HIIT under similar environmental conditions and time of the day (between 7 and 9 pm). Laboratory equipment and procedure familiarization were performed in a prior separate visit. The study was ethically approved by the institutional review board (IRB-2016-050) of the University of Taipei. All participants completed a written informed consent form and were fully explained of the experimental procedures. All procedures conformed to the Declaration of Helsinki for human participant research involvement.
Experiment Procedures
All participants were assessed at baseline, immediately before commencing the HIIT, and after the completion of the training intervention. Baseline preintervention and postintervention exercise assessments were conducted on separate visits after at least 3 d of recovery.
Baseline assessment
Baseline assessments included anthropometry, body composition, blood pressure, glycemic control, aerobic and anaerobic capacity, and hormone responses. Anthropometric measurements included height, weight, and waist and hip circumferences. Body composition was measured using dual-energy x-ray absorptiometry (DEXA; GE Healthcare Lunar iDXA, Madison, WI) and was analyzed for BF%, fat mass, and muscle mass. Total and regional body fat measured by DEXA has been validated in women after an exercise intervention as previously reported (7), and all DEXA measurements were performed by the same qualified DEXA technician. Blood pressure (both systolic and diastolic) was measured using a wrist cuff (Omron, Kyoto, Japan) two times, and the average was recorded.
Oral glucose tolerance test
Baseline oral glucose tolerance test (OGTT) involved participants consuming 75 g of glucose dissolved into 500 mL of water after ≥10 h of overnight fast. Capillary blood samples were collected at 0, 30, 60, 90, and 120 min, and immediately analyzed for glucose levels using a portable analyzer (Accu-chek; Roche Diabetes Care, Indianapolis, IN).
Hormones, endotoxins, and lipid profile
Venous blood was collected in the morning after an overnight fast using a vacutainer. The sample (5 mL) was then centrifuged at 3000 rpm for 10 min, and the serum supernatant was stored in a refrigerator at −80°C. Serum plasma cortisol and testosterone were both analyzed by an enzyme-linked immunosorbent assay (ELISA) reader (Infinite M200 Pro; Tecan Group Ltd, Mannedorf, Switzerland) using ELISA assay kits for cortisol and testosterone (IBL America, Minneapolis, MN). Serum plasma endotoxins (lipopolysaccharides) were analyzed using an endotoxin assay kit (ToxinSensor Chromogenic LAL; Genscript, Piscataway, NJ) for concentration based on the limulus amebocyte lysate method (23). HDL cholesterol and triglycerides (TG) were analyzed using an automated biochemical analyzer (AU400 Olympus, Tokyo, Japan).
For insulin levels, blood samples (200 μL) were collected into test tubes (Eppindorf, Hamburg, Germany), placed on ice and centrifuged (10 min at 4000 rpm; ThermoFisher, Waltham, MA), and analyzed for plasma insulin using an ELISA insulin assay kit (Mercodia, Uppsala, Sweden). Homeostasis model of assessment for insulin resistance (HOMA-IR), quantitative insulin sensitivity check index (QUICKI), and Matsuda–Defronzo index for insulin sensitivity (Matsuda index) were calculated from fasting and OGTT glucose and insulin data as described previously (24–26).
Cardiorespiratory fitness
Participants performed an incremental ramp cycling exercise test protocol using an ergometer (Monark 839E, Varberg, Sweden). Oxygen uptake (V˙O2) and carbon dioxide (V˙CO2), and heart rate (HR) were continuously measured throughout the test using an online gas analyzer (MetaMax 3B; Cortex Biophysik, Leipzig, Germany). The test started with a 5-min warm-up at 50 W, followed by a power output increase of 1 W every 2 s at 60 rpm until volitional exhaustion, which was defined by meeting two of the following criteria: a) the cadence falling below 50 rpm, b) V˙O2 leveling off, and c) respiratory exchange ratio ≥1.05. Peak oxygen uptake (V˙O2peak) was defined as the highest V˙O2 attained over the final 15 s of the test.
Anaerobic capacity
An all-out 30-s Wingate anaerobic test was performed on a cycling ergometer (Cyclus2, Leipzig, Germany). Saddle, handlebar height and distance, and toe clip were all adjusted. Participants warmed up for 3 min and maximally cycled for 30 s with a load of 0.075 kilopond per kilogram of body mass. Anaerobic capacity measurements included peak power (PP), mean power (MP) outputs relative to body mass (average 5-s intervals within a 30-s anaerobic exercise test) and fatigue index (FI) calculated as FI = [(PP − minimum power)/PP × 100].
Supplementation protocol
The CAF group was supplemented with 3 mg·kg−1·bw−1 of anhydrous caffeine (Sigma-Aldrich, St Louis, MO), and the PLA group received 3 mg·kg−1·bw−1 of placebo tapioca starch 1 h before each HIIT session. Capsules were of similar appearance and taste, and none of the participants knew the capsules’ content.
HIIT training protocol
Both CAF and PLA performed similar HIIT exercise, three HIIT sessions per week for 8 wk. Each training session consisted of 10 sets of 60-s sprints, interspersed by a 60-s dynamic recovery (at 50 W) on an ergometer (Ergoline, Bitz, Germany). The training intensity corresponding to 90% of maximal HR (HRmax) was individually calculated from the HRmax obtained during the initial cardiorespiratory V˙O2peak test, and the corresponding pedaling rate was 80–100 rpm. HR was monitored throughout each training session, and the range of 85%–95% HRmax was considered an acceptable training intensity.
Postintervention measurements
All baseline assessments of body composition, fasting plasma glucose, fasting insulin, postprandial OGTT, cardiorespiratory fitness, and anaerobic capacity were repeated at the completion of the 8-wk exercise training intervention. Postintervention assessments, including OGTT, were all performed 1 wk (7.5 ± 0.5 d) after the last training bout was completed. Anaerobic capacity and cardiorespiratory assessments were also separated by at least 3 d.
Data analysis and statistics
All data were presented as mean ± SEM. Two-way repeated-measures mixed ANOVA was used to compare the training effects of HIIT, pre versus post as within-factor, and the supplement effects, CAF versus PLA as between-factor. Tukey post hoc test was used to detect the specific time course differences and where relevant paired t-test was used. For correlations, Pearson’s product moment correlation coefficient test was used. Effect size was calculated using Cohen’s d statistics. Area under the curve (AUC) was calculated using the trapezoidal method. For all statistics, IBM SPSS statistics version 20 was used and significance level was set at P < 0.05.
RESULTS
Data are summarized in Table 1. HIIT decreased BF% in both PLA (P < 0.001, 95% confidence interval (CI), 42.2–37.5) and CAF (P < 0.001; 95% CI, 41.1–36.2), but there was no interaction between HIIT and CAF intake. HIIT decreased fat mass in both PLA (P < 0.01; 95% CI, 31.5–24.0) and CAF conditions (P < 0.01; 95% CI, 30.8–23.2), with no interaction effect. Muscle mass increased in PLA (P < 0.01; 95% CI, 38.4–43.1) and CAF conditions (P < 0.01; 95% CI, 39.2–43.9), with no interaction effect. There were no significant effects of HIIT, CAF intake, or interaction on BMI or waist-to-hip ratio on the lipid profile of neither HDL nor TG (Table 1).
;)
TABLE 1: Participants’ characteristics before (pre) and after (post) HIIT intervention with either placebo (PLA) or caffeine (CAF) analyzed by two-way mixed ANOVA.
OGTT glucose showed a significant interaction between HIIT and CAF supplementation (P = 0.001, ANOVA main effects). Postprandial OGTT glucose was significantly increased after training in the PLA group (P = 0.021), especially at 90 min after OGTT (P = 0.049), but was decreased in the CAF group (P = 0.019), especially at 120 min (P = 0.024; Fig. 1). These disparate HIIT-alone versus HIIT with CAF effects on postprandial glucose OGTT were irrespective of baseline fasting glucose, which was not different between the CAF and PLA groups. OGTT insulin also showed a significant interaction between HIIT and CAF supplementation (P = 0.019). OGTT insulin in the PLA group was increased (P = 0.008), especially at min 60 (P = 0.038), but it remained almost unchanged in the CAF group (P = 0.924; Fig. 1). These HIIT effects were also irrespective of baseline fasting insulin, which was not different between the CAF and PLA groups. Between-group (PLA vs CAF) effects were found for insulin (P = 0.029, ANOVA main effects) but not for glucose (P = 0.744, ANOVA main effects).
FIGURE 1: Blood glucose and insulin responses to 2-h OGTT after HIIT with either placebo (PLA) or caffeine (CAF) supplementation. HIIT had significant effects on glucose (ANOVA main within-group training effects, P < 0.05) and interacted with both OGTT glucose and OGTT insulin responses (P < 0.05). After HIIT, OGTT glucose and OGTT insulin responses were elevated in the PLA condition (P < 0.05) but remained unchanged in the CAF condition. Between-group (PLA vs CAF) effects were found for OGTT insulin (P < 0.05, ANOVA main effects) but not for OGTT glucose. $Significantly higher post-HIIT-PLA than pre-HIIT-PLA. #Significantly lower post-HIIT-CAF than pre-HIIT-CAF. d, Cohen’s effect size d; Pre-CAF, pre-HIIT in the CAF condition; Pre-PLA, pre-HIIT in the PLA condition; Post-CAF, post-HIIT in the CAF condition; Post-PLA, post-HIIT in the PLA condition. Data presented as means ± SEM (n = 24, 12 in each group).
The AUC for OGTT glucose response showed a significant interaction between HIIT and caffeine supplementation (P = 0.041). Independent post hoc group analysis showed that HIIT induced a 14.5% higher glucose AUC in the PLA treatment (P = 0.08; effect size, d = 0.28), but a 19.1% lower glucose AUC in the CAF treatment (P = 0.19; effect size, d = 0.51). No main effects of HIIT or interaction with caffeine were found for AUC in insulin. Independent post hoc group analysis showed a 28.3% increase in insulin AUC in the PLA group after HIIT (P = 0.08; effect size, d = 0.59), whereas there was almost no change (0.04%) in the CAF group after HIIT (P = 0.998; effect size, d = 0.001).
HOMA-IR, QUICKI, and Matsuda indices were not significantly affected by HIIT alone and HIIT with CAF, nor did they show any training/supplement interaction effects (Table 1).
Endotoxin levels were not significantly affected by training in both the CAF and PLA groups, nor was there any interaction between the two groups. However, independent post hoc group analysis showed a 30.8% increase in endotoxins for the PLA group after HIIT (P = 0.07; effect size, d = 0.78; 95% CI, 5.7–8.7), whereas there was almost no change (0.2%) in endotoxins in the CAF group after HIIT (effect size, d = 0.003; 95% CI, 6.5–10.6; Fig. 2).
FIGURE 2: Endotoxins (lipopolysaccharides) response to CAF and PLA. HIIT induced a 32% increase in endotoxins (P = 0.07; effect size, d = 0.78) in PLA condition. Such effect was absent when CAF was supplemented (P = 0.99; effect size, d = 0.003). Data presented as means ± SEM (n = 24, 12 in each group).
Testosterone levels were decreased by HIIT in both CAF- and PLA-supplemented conditions (P = 0.005, ANOVA main effects), but there was no interaction effect (Table 1). No difference or interaction was found for cortisol levels, nor was there any difference or interaction for testosterone/cortisol ratio. However, average testosterone reduction (16%, pooled change; Table 1) was associated with pretraining cortisol (r = 0.44, P < 0.05) and insulin AUC (r = 0.47, P < 0.05), and with posttraining HOMA-IR (r = 0.44, P < 0.05). The correlation between pretraining cortisol and pretraining testosterone was stronger in the CAF group (r = 0.59, P < 0.05) than in the PLA group (r = 0.44, P = 0.15), but those correlations were negated after training. Furthermore, a strong negative correlation was found for CAF between the HIIT-induced relative change of testosterone and HOMA-IR (r = −0.61, P < 0.05) compared with nonsignificant weak correlation for the PLA group (r = −0.23).
Cardiorespiratory capacity indicated by V˙O2peak was significantly increased by HIIT (ANOVA main effect, P < 0.001) in both the CAF and PLA groups, but there was no interaction effect (Table 1). Anaerobic capacity indicated by PP and MP was also significantly increased after HIIT training in both the PLA and CAF groups (ANOVA main effects, P = 0.003 and P < 0.001, respectively; Table 1), but neither PP nor MP showed any interaction effects between HIIT and CAF.
DISCUSSION
The results of the study confirm fat loss effect of HIIT with decreased testosterone level in women with obesity. Here we also provide novel evidence for the benefit of caffeine supplementation on the side effects of HIIT of increased insulinemia and endotoxemia. This study advances obesity lifestyle prevention guidelines, especially how HIIT exercise intervention and nutritional components interact to reverse obesity-induced risks of metabolic endotoxemia, hyperglycemia, hyperinsulinemia, and associated androgenic, cardiometabolic, and body composition obesity characteristics. In women with obesity, the study is first to show concurrent amelioration of obesity-induced endotoxicity, hyperglycemia, and hyperinsulinemia after HIIT intervention, especially when caffeine intake is combined with HIIT.
The present study found disparate HIIT effects on postprandial glucose and insulin responses between CAF and PLA treatments, which were independent of their fasting levels. HIIT alone increased both OGTT glucose and insulin levels, whereas caffeine reduced them, which suggests that caffeine ameliorated the HIIT-induced elevation in glucose–insulin responses (Fig. 1). The observed between-group (PLA vs CAF) effects found for insulin but not for glucose suggest a better insulin-lowering effects of caffeine, which dissipated any HIIT-induced elevation of insulin, acting as a regulator for glucose response. Chronic consumption of caffeine has long been shown to reduce diabetes risk in women, and a 30% reduced risk was found with an intake of three daily cups of coffee (21), which is close to a caffeine dose of 3 mg·kg−1 bw−1 provided to women with obesity in this intervention. Independent mechanisms of chronic caffeine intake in animals and humans have favored a potential caffeine action on glucose uptake and insulin sensitivity (27,28). For example, caffeine was associated with a potential increase in adiponectin regulatory effects on decreasing glucose and improving insulin sensitivity in men and women with diabetes (27). Treatment of caffeine for 15 d in high-fat–fed rats prevented the development of insulin resistance and reversed the increased weight gain and visceral fat mass (28). Chronic caffeine effects in the latter study were explained by α- and β-adrenergic receptor antagonist-related decrease in circulating catecholamines (28). Such effects may explain a better response in glucose and insulin responses, and a better insulin resistance when HIIT was combined with CAF compared with HIIT alone in women with obesity.
HIIT-induced adaptations of glucose and insulin are not well understood in women with obesity. Previously reported effects of HIIT on postprandial glucose and insulin responses were equivocal, perhaps due to insufficient intervention duration for the HIIT adaptations to occur, which affects making comparisons with a longer duration used in the present data. For example, a suggested decrease in 2-h postprandial insulin and glucose levels was reported after only a single bout of intense exercise at 80% V˙O2peak (29), whereas an improvement in insulin index was reported after only six HIIT sessions within 2 wk of training in adults with early diabetes (30). Adaptive short-term HIIT effects on glucose were explained by a concomitant increase in mitochondrial capacity (citrate synthase maximal activity, protein complex and subunits, and GLUT4) (31). However, longer-term HIIT interventions are lacking to confirm such mechanisms. In a recent study with a similar duration but different design compared with those used in the present study (32), HIIT combined with four different dietary regimes showed that insulin increased while glucose decreased in response to diet and HIIT, whereas both remained unchanged with two of the diets used, which suggest a nonuniform glucose–insulin response to HIIT–nutritional combination. Therefore, it is reasonable to attribute a combined HIIT with caffeine effects to synergistically induced better glycemic and insulin response compared with HIIT alone.
The present study is the first to demonstrate chronic HIIT-induced side effects on endotoxins in women with obesity. The phenomenon of leaky gut endotoxins (e.g., lipopolysaccharides) passed through the outer membrane of the intestinal wall from the large intestine into the blood is a major cause of systemic inflammation, insulin resistance, fat accumulation, and obesity (6). Our results showed a HIIT-induced sizeable increase in endotoxins only in the PLA group (31.5% increase), which was attenuated by caffeine intake in the CAF group (Fig. 2). Protective anti-inflammatory mechanisms against endotoxins have only been found in response to moderate-intensity exercise training, including an increase in foreign-body phagocytic capacity (e.g., increased Kupffer cell number), which triggers an attenuation in endotoxin-induced inflammatory cytokine levels (33). However, such positive effects are reversed in response to heavy and severe exercise intensities, possibly due to the heat stress and oxidative damage during intense exercise causing disruption to intestinal epithelial cell tight junction proteins and resulting in increased permeability to luminal endotoxins (33). Compared with moderate exercise intensity (90% of first ventilatory threshold (VT)), heavy (between first and second VT) and severe exercise intensities (midpoint between second VT and V˙O2peak), which are close to this study’s HIIT intensity range of 90% HRmax, have been shown to acutely increase circulating endotoxins and cytokine responses due to higher endotoxin-stimulated TNF-α, IL-6/IL-10 ratio, and higher lactate levels (34,35). Such mechanisms may be explained by a possible depletion of Kupffer cell number and signaling by endotoxins binding to TLR-4 present on endothelial cells, macrophages, and monocytes, and adipose tissue inflammation (7,15).
Higher HIIT-induced endotoxins in PLA condition was accompanied by an elevated OGTT glucose response, requiring insulin action (Figs. 1, 2), which suggests blunted adaptations to systemic inflammation and metabolic responses to endotoxemia (35). It is likely that caffeine ameliorated glucose response in CAF condition and attenuated insulin response (Fig. 1), which enhanced HIIT adaptive response to endotoxins. However, further research is needed to ascertain the combined endotoxic-induced glucose and insulin responses and consequent inflammatory, cytokine endotoxic responses to exhaustive exercise such as HIIT in individuals with advanced obesity.
The HIIT-induced reduction (average 16%) in testosterone can be partly explained by prelevel cortisol and insulin (Table 1), which showed a positive moderate correlation with testosterone relative change. Although testosterone was reduced in both PLA and CAF conditions, its relative change showed a strong negative correlation with relative HOMA-IR change in the CAF (r = −0.61), but not in the PLA. This suggests that greater testosterone reduction in CAF versus PLA (17.5% vs 14%) may partly explain better insulin sensitivity found in the CAF group. In premenopausal and perimenopausal women, higher free testosterone was associated with higher insulin, glucose, and HOMA-IR (13). Reported associations in women with diabetes have also documented higher levels of free testosterone and low levels of sex hormone–binding globulin (SHBG), which contrasts those found in diabetic men who demonstrate low levels in both testosterone and SHBG (36). Such phenomenon in women with obesity was explained by testosterone altering the availability of estrogens, by competitively binding with SHBG, and/or by acting as an estrogen precursor (13,36). Furthermore, augmented androgen concentrations have been associated with increased insulin levels and diabetes, and women with hyperandrogenemia exhibit insulin resistance, increased inflammatory cytokines, dyslipidemia, and metabolic syndrome (37). HIIT-induced combined androgenicity effects seem to favor the concurrent CAF adaptations, which ameliorated HIIT-induced glucose–insulin response and reflected a favorable reduction in testosterone levels.
Habitual caffeine intake is often associated with fat loss through its thermogenic effects, stimulated fat oxidation, suppressed leptin, and appetite in women (38). Together with HIIT, there was a small but significant decrease in whole body fat (≃0.6–0.7 kg) in both the PLA and CAF groups, which agrees with previously reported HIIT-induced reduction in whole and regional fat mass (1–3 kg) reported in overweight women (8). However, higher initial adiposity in female participants in the present study (fat percentage, 40%) may have dissipated any added caffeine-induced fat loss responsiveness.
The present study confirms the cardioprotective benefits of HIIT in women with obesity irrespectively of caffeine intake, as indicated by ≃24% increase in V˙O2peak (Table 1), which is preventative of cardiovascular morbidity and mortality (39). Nonetheless, a moderate (≃11%) increased anaerobic capacity, coupled with a higher fatigue rate (≃30%), suggests a HIIT physical performance limitations, which should be considered when promoting HIIT as a public health strategy for sedentary women with obesity.
CONCLUSIONS
In women with obesity, HIIT combined with caffeine intake induces concurrent amelioration of obesity-induced endotoxicity, irregular androgenicity, hyperglycemia, and hyperinsulinemia. However, HIIT-alone induced hyperinsulinemia and increased endotoxicity, though within a healthy range, favors the synergetic HIIT–caffeine as a public health obesity prevention strategy.
This work was supported by a grant (No. 106-2410-H-845-024) from the Ministry of Science and Technology, Taipei, Taiwan. The authors declare no conflict of interest.
The results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation. The results of the present study do not constitute endorsement by the American College of Sports Medicine.
Author contributions: A. A. developed the study design and manuscript concept, analyzed the data, and wrote the full manuscript. M.-J. H. participated in data collection, analysis, and writing of the manuscript. C.-H. K. participated in critical editing of the manuscript. C.-W. H. designed the study, coordinated data collection, and participated in data analysis and writing of the manuscript. All authors have read and approved the final version of the manuscript.
REFERENCES
1. Bhurosy T, Jeewon R. Overweight and obesity epidemic in developing countries: a problem with diet, physical activity, or socioeconomic status?
ScientificWorldJournal. 2014;2014:964236.
2. Colditz GA, Willett WC, Rotnitzky A, Manson JE. Weight gain as a risk factor for clinical diabetes mellitus in women.
Ann Intern Med. 1995;122:481–6.
3. Bastard JP, Maachi M, Lagathu C, et al. Recent advances in the relationship between obesity, inflammation, and insulin resistance.
Eur Cytokine Netw. 2006;17:4–12.
4. Neves AL, Coelho J, Couto L, Leite-Moreira A, Roncon-Albuquerque R Jr. Metabolic endotoxemia: a molecular link between obesity and cardiovascular risk.
J Mol Endocrinol. 2013;51(2):R51–64.
5. Haffner SM. Sex hormones, obesity, fat distribution, type 2 diabetes and insulin resistance: epidemiological and clinical correlation.
Int J Obes Relat Metab Disord. 2000;24(2 Suppl):S56–8.
6. Cani PD, Amar J, Iglesias MA, et al. Metabolic endotoxemia initiates obesity and insulin resistance.
Diabetes. 2007;56:1761–72.
7. Mach N, Fuster-Botella D. Endurance exercise and gut microbiota: a review.
J Sport Health Sci. 2017;6:179–97.
8. Trapp EG, Chisholm DJ, Freund J, Boutcher S. The effects of high-intensity intermittent exercise training on fat loss and fasting insulin levels of young women.
Int J Obes (Lond). 2008;32:684–91.
9. Tuomilehto J, Lindstrom J, Eriksson JG, et al. Finnish Diabetes Prevention Study Group. Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance.
N Engl J Med. 2001;344:1343–50.
10. Gibala MJ, McGee SL. Metabolic adaptations to short-term high-intensity interval training: a little pain for a lot of gain?
Exerc Sport Sci Rev. 2008;36:58–63.
11. Gillen JB, Percival ME, Ludzki A, Tarnopolsky MA, Gibala MJ. Interval training in the fed or fasted state improves body composition and muscle oxidative capacity in overweight women.
Obesity. 2013;21:2249–55.
12. Biddle SJ, Batterham AM. High-intensity interval exercise training for public health: a big HIT or shall we HIT it on the head?
Int J Behav Nutr Phys Act. 2015;12:95.
13. Sutton-Tyrrell K, Wildman RP, Matthews KA, et al. Sex hormone–binding globulin and the free androgen index are related to cardiovascular risk factors in multiethnic premenopausal and perimenopausal women enrolled in the study of women across the nation (SWAN).
Circulation. 2005;111:1242–9.
14. Wang JM, Gong TY, Zhao ZJ. Effects of exercise on sex hormones and expression of relevant genes in the hypothalamus in obese mice.
J Biol Regul Homeost Agents. 2019;33:517–23.
15. Boutagy NE, McMillan RP, Frisard MI, Hulver MW. Metabolic endotoxemia with obesity: is it real and is it relevant?
Biochimie. 2016;124:11–20.
16. Alkhatib A, Tsang C, Tiss A, et al. Functional foods and lifestyle approaches for diabetes prevention and management.
Nutrients. 2017;9(4). pii: E1310.
17. Diepvens K, Westerterp KR, Westerterp-Plantenga MS. Obesity and thermogenesis related to the consumption of caffeine, ephedrine, capsaicin, and green tea.
Am J Physiol Regul Integr Comp Physiol. 2007;292:R77–85.
18. Alkhatib A, Seijo M, Larumbe E, Naclerio F. Acute effectiveness of a “fat-loss” product on substrate utilization, perception of hunger, mood state and rate of perceived exertion at rest and during exercise.
J Int Soc Sports Nutr. 2015;12:44.
19. Alkhatib A, Tuomilehto J. Lifestyle diabetes prevention. In: Huhtaniemi I, Martini L, editors.
Encyclopaedia of Endocrine Diseases. 2nd ed. Amsterdam (the Netherlands): Elsevier; 2018. pp. 148–59.
20. Shi X, Xue W, Liang S, Zhao J, Zhang X. Acute caffeine ingestion reduces insulin sensitivity in healthy subjects: a systematic review and meta-analysis.
Nutr J. 2016;15:103.
21. Tuomilehto J, Hu G, Bidel S, Lindström J, Jousilahti P. Coffee consumption and risk of type 2 diabetes mellitus among middle-aged Finnish men and women.
JAMA. 2004;291:1213–9.
22. Al-Attas OS, Al-Daghri NM, Al-Rubeaan K, et al. Changes in endotoxin levels in T2DM subjects on anti-diabetic therapies.
Cardiovasc Diabetol. 2009;8:20.
23. Ding JL, Ho B. Endotoxin detection—from limulus amebocyte lysate to recombinant factor C.
Subcell Biochem. 2010;53:187–208.
24. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man.
Diabetologia. 1985;28:412–9.
25. Matsuda M, DeFronzo RA. Insulin sensitivity indices obtained from oral glucose tolerance testing: comparison with the euglycemic insulin clamp.
Diabetes Care. 1999;22:1462–70.
26. Chen H, Sullivan G, Yue LQ, Katz A, Quon MJ. QUICKI is a useful index of insulin sensitivity in subjects with hypertension.
Am J Physiol Endocrinol Metab. 2003;284:E804–12.
27. Bhaktha G, Nayak BS, Mayya S, Shantaram M. Relationship of caffeine with adiponectin and blood sugar levels in subjects with and without diabetes.
J Clin Diagn Res. 2015;9:BC01–3.
28. Conde SV, Nunes da Silva T, Gonzalez C, Mota Carmo M, Monteiro EC, Guarino MP. Chronic caffeine intake decreases circulating catecholamines and prevents diet-induced insulin resistance and hypertension in rats.
Br J Nutr. 2012;107(1):86–95.
29. Francois ME, Little JP. Effectiveness and safety of high-intensity interval training in patients with type 2 diabetes.
Diabetes Spectr. 2015;28:39–44.
30. Shaban N, Kenno K, Milne K. The effects of a 2 week modified high intensity interval training program on the homeostatic model of insulin resistance (HOMA-IR) in adults with type 2 diabetes.
J Sports Med Phys Fitness. 2014;54:203–9.
31. Little JP, Gillen JB, Percival ME, et al. Low-volume high-intensity interval training reduces hyperglycemia and increases muscle mitochondrial capacity in patients with type 2 diabetes.
J Appl Physiol. 2011;111:1554–60.
32. Asle Mohammadi Zadeh M, Kargarfard M, Marandi SM, Habibi A. Diets along with interval training regimes improves inflammatory & anti-inflammatory condition in obesity with type 2 diabetes subjects.
J Diabetes Metab Disord. 2018;17(2):253–67.
33. Hurst SM, Lyall K, Hurst R, Stevenson L. Exercise-induced elevation in plasma oxidative generating capability augments the temporal inflammatory response stimulated by lipopolysaccharide.
Eur J Appl Physiol. 2009;107:61–72.
34. Antunes BM, Campos EZ, dos Santos RVT, et al. Anti-inflammatory response to acute exercise is related with intensity and physical fitness.
J Cell Biochem. 2019;120:5333–42.
35. Brix-Christensen V, Andersen SK, Andersen R, et al. Acute hyperinsulinemia restrains endotoxin-induced systemic inflammatory response. An experimental study in a porcine model.
Anesthesiology. 2004;100:861–70.
36. Andersson B, Märin P, Lissner L, Vermeulen A, Björntorp P. Testosterone concentrations in women and men with NIDDM.
Diabetes Care. 1994;17:405–11.
37. Azziz R, Carmina E, Dewailly D, et al. Criteria for defining polycystic ovary syndrome as a predominantly hyperandrogenic syndrome: an androgen excess society guideline.
J Clin Endocrinol Metab. 2006;91:4237–45.
38. Westerterp-Plantenga MS, Lejeune MP, Kovacs EM. Body weight loss and weight maintenance in relation to habitual caffeine intake and green tea supplementation.
Obes Res. 2005;13:1195–204.
39. Wisløff U, Coombes JS, Rognmo Ø. CrossTalk proposal: high intensity interval training does have a role in risk reduction or treatment of disease.
J Physiol. 2015;593:5215–7.
