Aston et al., 2010 [20]randomized, balanced, two-way crossover intervention;
UK12 overweight but otherwise healthy women, with a BMI of 25–30 kg/m² and aged 18–65 yearsMay–December 2006No differences were observed in glucose profiles between higher and lower GI interventions in the controlled or ad libitum setting.Babir et al., 2023 [32]randomized crossover trial;
Canada27 healthy, inactive participants, aged 18–35 years, inactive based on not meeting
the aerobic physical activity targets in the
Canadian 24 h Movement Guidelines for AdultsDecember 2021–May 2022Mean 24-h glucose after BWE and CON was not different. No differences were found between conditions for measures of glycemic variability or the postprandial glucose responses after ingestion of a 75 g glucose drink or lunch, dinner, and breakfast meals.Bermingham et al., 2022 [37]randomized single arm, single blinded multicenter intervention trial;
UK and US1102 healthy adults, aged 18–65 yearsJune 2018–May 2019Post-menopausal females had higher fasting blood measures compared with pre-menopausal females (p < 0.05 for all). Postprandial metabolic responses for glucose (2hiauc) and insulin (2hiauc) were higher (42% and 4% respectively) and CGM measures (glycemic variability and time in range) were unfavorable post- versus pre-menopause (p < 0.05 for all).Buscemi et al., 2013 [14]randomized controlled trial;
Italy40 participants, aged 18–60 years, BMI 25–49.9 kg/m², with ≥2 diagnostic criteria of metabolic syndrome, except T1DM or T2DMNovember 2010–February 2012The glycemic index of the diet significantly influenced the FMD (p < 0.005). In particular, the change in FMD was 2.3 ± 2.6% following the LGI diet, and −0.9 ± 3.6% after the HGI diet (p < 0.005). The mean 48-h glycemia decreased significantly after dietary treatment (p < 0.05), but no significant effect of the glycemic index of the diet on results was observed. The glycemic index of the diet significantly influenced the 48-h glycemic variability measured as coefficient of variability (CV%; p < 0.001). The CV% decreased after the LGI diet (from 23.5 to 20.0%) and increased after the HGI diet (from 23.6 to 26.6%).Buscemi et al., 2009 [36]observational, three group study;
Italy28 overweight or obese subjects (11 males, 17 females), aged 30–60 years, BMI 25–35 kg/m²The average CGM CV% increased from MS- group (21.1%) to the MS+ group (23.9%) and to the MS+/T2D group (27.4%), but it was not correlated with the CGM mean glycemia (r = 0.20; p = ns). Stepwise multiple correlation analysis showed that IL-6 predicted CGM CV% (R(2) = 0.35, beta = 0.13; p < 0.05) independently of BMI, waist circumference, adiponectin and insulin concentrations.Byun et al., 2020 [44]prospective, observational single center study;
South Korea23 patients, BMI < 35 kg/m², no diabetes, no cardiac disease, no other sleep disorderThose with moderate to severe OSA showed an increasing trend in blood glucose levels after sleep onset, whereas those without or with mild OSA showed a decreasing trend (F  = 8.933, p < 0.001).Castaldo et al., 2011 [39]observational study;
Italy79 non-diabetic Caucasian subjects, aged 23–70 years, with no history of cv diseaseIGT individuals had a worse cardiovascular risk profile, including higher IMT, and spent significantly more time in hypoglycemia than glucose-tolerant individuals.Hafiz et al., 2022 [16]randomized, crossover, controlled trial;
UK19 healthy adults aged 18–65 years, with fasting glucose < 5.6 mmol/L and BMI 18–29.9 kg/m²15 August–20 December 2019Postprandial glycemic responses were comparable between chickpea treatments, albeit significantly lower than the control (p < 0.001). All chickpea treatments elicited significantly lower C-peptide and GLP-1 responses compared to the control (p < 0.05), accompanied by enhanced subjective satiety responses (p < 0.05).Jansen et al., 2022 [24]randomized controlled trial;
US70 men and women aged 18–50 years, BMI ≥27 kg/m² and no known cardiovascular disease or diabetesMay 2018–May 2020
(stopped 2019 due to COVID-19 pandemic)Glucose metrics continued to decline after week 1 in the HC-Starch and HC-Sugar groups (p < 0.05), but not VLC. The number of participants with abnormal glucose tolerance by OGTT remained 10 (of 16) in VLC at start and end, but decreased from 17 to 9 (of 25) in both high-carbohydrate groups.Kang et al., 2013 [18]cross-sectional study;
China78 subjects with NGT and 55 IGR individuals 20–70 years of age, non-obese with BMI 18.5–24.9 kg/m² without heart, lung, liver or kidney diseaseThe postprandial fluctuations of glucose increased gradually with increased proportions of carbohydrate in breakfast in both IGR and NGT subjects. Compared with the NGT subjects for the HC meal, the IGR subjects consuming the MC meal had greater PGS, range of glucose concentrations, SD, and PPGE (p < 0.05).Kobayashi et al., 2013 [22]randomized crossover trial with 2 experimental conditions;
Japan8 male young adults, free from pathological conditions, with no medication/supplementsBreakfast skipping did not affect 24-h energy expenditure, fat oxidation, or thermic effect of food, but increased the overall 24-h average blood glucose (83 ± 3 vs. 89 ± 2 mg/dL, p < 0.05).Kuroda et al., 2015 [40]prospective study;
Japan72 patients, referred for PCI for CAD, 20–80 years of age under adequate treatment for dyslipidemia, with LDL levels <120 mg/dL under statin administration or <100 mg/dL under other treatment for dyslipidemiaJune 2012–April 2015LI was stepwisely increased according to the tertile of MAGE (1958 ± 974 [tertile 1] vs. 2653 ± 1400 [tertile 2] vs. 4362 ± 1858 [tertile 3], p < 0.001), whereas FCT was the thinnest in the tertile 3 (157.3 ± 73.0 μm vs. 104.0 ± 64.1 μm vs. 83.1 ± 34.7 μm, p < 0.001, respectively).
MAGE had the strongest effect on LI and FCT (standardized coefficient β = 0.527 and −0.392, respectively, both p < 0.001). Multiple logistic analysis identified MAGE as the only independent predictor of the presence of TCFA (odds ratio 1.034; p < 0.001).Li et al., 2018 [17]randomized controlled cross-over study;
UK37 healthy overweight (BMI > 25 kg/m²), non- smoking male volunteers, aged 36–70 years, with no known history of CVD or T2DM, and no medicationAugust 2016–February 2017After 4 weeks of intervention, blood glucose and low-density lipoprotein (LDL) cholesterol were significantly lower than baseline in both groups, but there was no difference between quinoa and control.
The cumulative area under the blood glucose curve for the last 4 days of the quinoa intervention tended to be lower than the first 4 days of washout (p = 0.054), and was significantly lower than the corresponding period of the wheat treatment (p = 0.039).Liao et al., 2020 [45]observational study;
US19 individuals (men and women), aged 18- 65 years, BMI ≥ 25 kg/m², engaged in less than 150 minutes of moderate-intensity physical activity per week in the previous monthThe summary acceptability scores for the self-monitoring period were 4.46 for CGM and 4.51 for Fitbit. Participants reported a significant decrease in the precontemplation stage and an increase in the action stage (p < 0.05).Little et al., 2014 [26]randomized, counter-balanced fashion;
Canada10 overweight or obese individuals (BMI >25 kg/m², 8 females/ 2 males) who were inactive (<2 bouts of exercise lasting a minimum of 30 minutes per bout per week)Exercise did not affect the PPG responses to lunch, but performing both HIIT and CMI in the morning significantly reduced the PPG incremental area under the curve (AUC) following dinner when compared with control (HIIT = 110 ± 35, CMI = 125 ± 34, control = 162 ± 46 mmol/L × 2 h, p < 0.05). The PPG AUC (HIIT = 125 ± 53, CMI = 186 ± 55, control =194 ± 96 mmol/L × 2 h) and the PPG spike (HIIT = Δ2.1 ± 0.9, CMI = Δ3.0 ± 0.9, control = Δ3.0 ± 1.5 mmol/L) following breakfast on the following day were significantly lower following HIIT compared with both CMI and control (p < 0.05). Absolute AUC and absolute glucose spikes were not different between HIIT, CMI, or control for any meal (p > 0.05 for all).Mikus et al., 2012 [29]observational cross sectional study;
US12 healthy volunteers (8 men, 4 women); 20–35 years of age, generally healthy (determined by detailed medical history questionnaire), recreationally active (≥ 10.000 pedometer steps per day)During 3 d of reduced physical activity (12,956 ± 769 to 4319 ± 256 steps per day), PPG increased at 30 and 60 min after a meal (6.31 ± 0.19 to 6.68 ± 0.23 mmol/L and 5.75 ± 0.16 to 6.26 ± 0.28 mmol/L, p < 0.05 relative to corresponding active time point), and ΔPPG increased by 42%, 97%, and 33% at 30, 60, and 90 min after a meal, respectively (p < 0.05). Insulin and C-peptide responses to the OGTT increased after 3 d of reduced activity (p < 0.05), and the glucose response to the OGTT did not change significantly.Nakayama et al., 2022 [27]randomized crossover trial;
Japan12 healthy young adult males, with postprandial hyperglycemia defined as blood glucose level >140 mg/dL at 30 or 45 min after a mealThe glucose concentrations after the meal were significantly lower in the home-based HIIE and MICE conditions than in the control condition (p < 0.001). There were no significant differences in the glucose concentrations between the home-based HIIE and MICE conditions.Parr et al., 2018 [19]randomized, crossover study;
Australia13 men and women aged 40–70 years with overweight/obesity, sedentary and lifestyles with IFG and/or IGTFebruary–September 2016Total glucose area under the curve (AUC; +5.7 mmol/L/h, p = 0.019) and mean plasma glucose concentrations (+0.5 mmol/L, p = 0.014) were greater after HE-BF compared to LE-BF. In the HE-BF condition, compared to LE-BF, there was a greater incremental area under the curve (iAUC) for plasma glucose post-breakfast (+44 ± 59%, p = 0.007), but lower iAUC post-lunch (–55 ± 36%, p < 0.001). Total insulin AUC was greater (+480 mIU/mL/h, p < 0.01) after HE-BF compared to LE-BF. Twenty-four-hour (24 h) CGM revealed no differences in mean glucose and total AUC between conditions.Philippou et al., 2008 [15]randomized parallel group trial;
UK13 subjects, aged 35–65 years and one heart disease risk factor (BMI 27–35 kg/m², waist ≥ 88 cm females or ≥94 cm males, cholesterol: HDL ratio ≥ 5.0 mmol/L, BP systolic > 130 mmHg or diastolic > 85 mmHg)A significantly different dietary GI was achieved in the low GI (median: 51.3 (IQR: 51.0–52.0) compared to the high GI (59.3 (59.2–64.0) (p = 0.032) group. By week 12, the low GI group also had a significantly lower 24-h area under the curve (AUC) (7556 (7315–8434) vs. 8841 (8424–8846) mmol-h/L (p = 0.045) and overnight AUC (2429 (2423–2714) vs. 3000 (2805–3072) mmol-h/L (p = 0.006) glucose as measured by CGMS.Reynolds et al., 2023 [31]randomized, crossover, exploratory study;
US20 participants (11 old, 9 younger), aged >55 years or between 18–40 years completing at least 30 min of moderate to vigorous exercise on 3 or more days per week during the past 3 monthsSignificant main effects of age (p = 0.002) and time (p  <  0.001) existed for 1-h PPIG, but no effect of phase or interactions was found (p  >  0.05). Significant main effects (p  <  0.05) of age (old: 114  ±  1 mg/dL, young: 106  ±  1 mg/dL), phase (NOEX: 112  ±  1 mg/dL, EX: 108  ±  1 mg/dL), and time (0 min: 100  ±  2, 30 min: 118  ±  2, 60 min: 116  ±  2, 90 min: 111  ±  2, 120 min: 108  ±  2 mg/dL) in 2-h PPIG were detected, but no interaction was found (p  >  0.05). Only significant main effects of phase (NOEX: 14  ±  1 and EX: 12  ±  1, p  >  0.05) were found for 24-h blood glucose standard deviation.Rothberg et al., 2016 [42]cross-sectional, observation study;
Australia32 subjects with DM, and 31 subjects without diabetes (without history of cv events)Low-frequency (LF) power, high-frequency (HF) power, and total power (TP) of HRV were negatively associated with BGL in participants with DM. The ratio of LF to HF was positively correlated with BGL.Salkind et al., 2014 [35]uncontrolled, observational study;
US36 morbidly obese (BMI ≥ 40 kg/m²) applicants of the TV show The Biggest LoserMorbidly obese prediabetic subjects (n = 15) had GV metrics indistinguishable from those morbidly obese subjects who were normoglycemic. Normoglycemic and prediabetic morbidly obese individuals have higher GV compared with normal weight, nondiabetic individuals.Selvin et al., 2021 [38]pilot study of a prospective epidemiology study (ARIC);
US27 adults (8 with T2DM, 19 without diabetes)October–November 2019In persons without diabetes, there was a wide range of CGM parameters: range of mean glucose, 83.7–124.5 mg/dL, SD 12.2–27.3 mg/dL, CV 14.0–26.7%, and TIR 71.1–99.5%. In persons with diabetes, the range of mean CGM glucose was 105.5–223.0 mg/dL, SD, 22.3–86.6 mg/dL, CV 18.2–38.8%, TIR 38.7–98.3%. There was a high prevalence of hypoglycemia (glucose < 70 or <54 mg/dL) in both persons with and without diabetes.Sezer et al., 2021 [43]cross-sectional study;
Turkey27 non-diabetic, normotensive healthy subjects age over 18 years, BMI < 25 kg/m² and absence of chronic disease (obesity, DM, lipid abnormalities, hypertension, liver, kidney and cv disease)In the correlation analysis between glycemic variability parameters and BPV parameters, SD of 24-h SBP was correlated with the SD of MBG (r = 0.52, p = 0.006), MAGE (r = 0.49, p = 0.009), and MODD (r = 0.46, p = 0.015). SD of daytime SBP was correlated with, MAGE (r = 0.42, p = 0.03) and MODD (r = 0.43, p = 0.02).
There is a correlation between glycemic variability and BPV variables in normoglycemic and normotensive healthy individuals.Smith et al., 2021 [28]parallel randomized controlled trial;
Sweden16 adults with obesity, self-perceived sedentary lifestyle, a sedentary occupation or unemployment, an age between 18–60 years and BMI 30–45 kg/m²1 February 2017–31 July 2019Mean (±SD) fasting glucose levels [−0.34 (±0.37) mmol/L] and daily glucose variation [%CV; −2% (±2.2%)] were reduced in FABS, suggesting a modest benefit for glycemic control that was most robust at higher volumes of daily activity.Solomon et al., 2020 [33]randomized, counter-balanced, controlled trial;
UK48 participants, aged 18–65 years, BMI 18–30 kg/m², generally healthy and physically activeOctober 2017–November 2018Walking and bodyweight exercises immediately after the meal improved mean, CV, and AUC glucose (p ≤ 0.05 vs. control), while standing immediately after the meal only improved AUC glucose (p ≤ 0.05 vs. control) and nearly improved mean glucose (p =  0.06).
Mean, CV, and AUC glucose were not affected by standing, walking, or bodyweight exercise conducted immediately before, or 30 min after the meal (all p >  0.05 vs. control).Teshima et al., 2020 [41]observational two group, cohort study;
Japan18 patients with enlargement of cardiac silhouette CTR > 50%, NYHA II-IV and impaired systolic/diastolic function (EF < 50% or E/e > 14)The average level and peak value of interstitial glucose concentrations, and the duration of hyperglycemia (≥140 mg/ dL) were not significantly different between Heart failure (+) and Heart failure (−). The duration of hypoglycemia (<80 mg/dL) was significantly longer and the trough value was significantly lower in Heart failure (+) compared with Heart failure (−).Tey et al., 2017 [21]randomized, crossover trial;
Singapore10 males, healthy, aged 21–50 years, BMI 18.5–25 kg/m²No significant differences were found in mean 24-h glucose, iAUC and total AUC for glucose, and 24-h glycemic variability between the four test beverages. Twenty-four-hour glucose profiles did not differ between beverages sweetened with non-nutritive (artificial vs. natural) and nutritive sweeteners.Timmer et al., 2022 [23]randomized, three way, crossover trial;
Netherlands24 participants, aged 18–65 years, BMI 18.5–30 kg/m², no reported health problems, no state of (pre)diabetes15–24 February 2021
8–17 March 2021The highest glycemic excursions to a standardized carbohydrate-rich snack (198 kcal) were observed in the morning, while a more dampened but prolonged response was observed in the evening.Trim et al., 2023 [30]randomized, controlled trial;
UK20 healthy, young (20–45 years old) males; selected by medical and psychological screeningFollowing long-term bed rest, fasting plasma insulin concentration increased 40% (p = 0.004) and glucose disposal during the HIEC decreased 24% (p < 0.001). Interstitial daily glucose total area under the curve (tAUC) from pre-to post-bed rest increased on average by 6% (p = 0.041), despite a 20 and 25% reduction in total caloric and carbohydrate intake, respectively.
The nocturnal period (00:00–06:00) showed the greatest change to glycemia with glucose tAUC for this period increasing by 9% (p = 0.005). CGMS measures of daily glycemic variability (SD, J-Index, M- value and MAG) were not changed during bed rest.Zhang et al., 2023 [25]randomized, crossover trial;
China20 young adults (11 males, 9 females) with obesity (BMI ≥ 25 kg/m²), aged 18–35 years, who were sedentary (self-reported sedentary time > 8 h/day) and insufficiently active (<150 min moderate-intensity and/or 75 min vigorous intensity PA per week over the past 3 months)The 4-h PPG incremental area under the curve (iAUC) was 12.1% ± 30.9% and 21.5% ± 21.5% smaller after CONT (p = 0.022) and ACCU (p < 0.001), respectively, than after SIT. PPG concentrations were lower during CONT at 30–60 min and during ACCU at 30–105 min after breakfast than during SIT (all p < 0.05). The 4-h plasma insulin and C-peptide iAUC, and mean amplitude of glycemic excursions were lower after CONT and ACCU than after SIT (all p < 0.05).Zhang et al., 2021 [34]randomized, controlled crossover trial;
China20 male participants, overweight or obese, BMI ≥ 23 kg/m², aged 18–35 years, self-reported daily sedentary time > 8 h, and insufficient physical activity (measured by the Chinese version of the IPAQ Short Form)Compared with SIT, the 4-h incremental AUCs (iAUCs) for plasma PPG (−0.6 mmol·L⁻¹·h; p = 0.047) and insulin (−28.7%, p < 0.001) were reduced in 20 iP only, and C-peptide concentrations were lower after iP (−14.9%, p = 0.001) and 20 iP (−28.7%, p < 0.001). PPG reductions due to iP and 20 iP occurred only in men with a BMI > 27.5 kg/m² (iP, −11.2%; 20 iP, −14.7%; p = 0.047) and higher glucose iAUC values during SIT (iP, −25.5%; 20 iP, −25.7%; p < 0.001).