Design

This was a 12-week, single-center randomized controlled trial conducted between 2 June 2021 and 30 November 2022. Ethical approval was obtained from the Institutional Ethics Committee (IRB/COHS/FAC/20/June2021), and written informed consent was provided by all participants prior to enrolment. The trial was prospectively registered with the Clinical Trials Registry of India (CTRI/2022/04/041762). The study was conducted and reported in accordance with the Consolidated Standards of Reporting Trials (CONSORT) guidelines for randomised controlled trials, including the updated 2025 recommendations [26].

Setting

The exercise intervention was performed at a tertiary care hospital in Ajman, United Arab Emirates (UAE).

Participants

Eligible participants included individuals who had been diagnosed with T2DM and were currently receiving treatment with oral hypoglycaemic agents with or without insulin therapy. The age range for inclusion was 30–65 years, and both males and females were eligible to participate. Participants meeting any of the following criteria were excluded from the study: confirmed respiratory disease, coronary artery disease, neurological disorders, musculoskeletal problems that could hinder exercise training, uncontrolled hypertension (systolic blood pressure > 180 mmHg or diastolic blood pressure > 120 mmHg), pregnancy, thyroid disorders, or lack of willingness to participate in the study. The purpose of the study and the benefits of participation were explained to the prospective participants. The participants were asked to read an information sheet and sign a written informed consent form.

A total of 200 participants underwent screening, and 90 eligible participants were ultimately recruited based on meeting the inclusion and exclusion criteria. Initial screening involved the use of the Physical Activity Readiness Questionnaire (PAR-Q) [27] and American Heart Association (AHA)/American College of Sports Medicine (ACSM) risk stratification questionnaire [28]. Eligible participants were categorized into a combined aerobic and resistance exercise training (A + R) group (A) (n = 30), a HIIT programme group (B) (n = 30), and a control group (C) (n = 30) (Fig 1) [26].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. This is the Fig 1 Consort flowchart of participants.

https://doi.org/10.1371/journal.pone.0336898.g001

Randomization

Following evaluation by a physiotherapist, participants were randomly allocated to one of three groups: combined aerobic and resistance training (A + R), high-intensity interval training (HIIT), or control. The A + R group undertook an individually tailored, structured, and supervised programme combining aerobic and resistance exercises. The HIIT group performed structured high-intensity interval sessions, while the control group continued with standard care without structured exercise. Block randomization was employed using Sequentially Numbered Opaque Sealed Envelopes (SNOSE), with the principal investigator generating the allocation sequence and assigning participants according to the envelope draw. Assessments and exercise prescriptions were provided free of charge to all participants. The study used nine blocks, with 10 participants per block.

Intervention

Group A participated in a 12-week intervention that included an individualised combined aerobic and resistance exercise training program along with standard hospital care. Group B underwent a HIIT programme program adapted from the model described by Weston et al. [29], which typically involved repeated short intervals of high-intensity exercise interspersed with recovery periods. To suit the safety and physical capacity of participants with T2DM, the protocol was modified as follows: (1) the intensity was adjusted to 85–90% of peak heart rate rather than maximal effort; (2) interval durations were standardized to 1-minute high-intensity bouts followed by 1-minute active recovery, repeated for 10 cycles per session; and (3) a longer warm-up and cool-down phase (5 minutes each at 50–60% HRmax) was added. These modifications aimed to optimize safety, adherence, and tolerability while maintaining the physiological demands consistent with HIIT principles. Both programs incorporated activities such as walking and exercises for the upper and lower limbs, tailored to each participant based on their baseline six-minute walk distance (6MWD) and rating of perceived exertion (RPE). Exercises were demonstrated to the patients and their caregivers during the initial visit, with a patient education manual provided for guidance.

For the first two weeks, the participants engaged in supervised sessions within the hospital. Following the initial supervised sessions, participants transitioned to a structured home-based HIIT programme with predefined progressions implemented at weeks 2 and 6. Each session began with a 10-minute warm-up of light aerobic exercises, followed by interval walking alternating between 1-minute high-intensity bouts (at an RPE of 7–8 out of 10) and 1-minute active recovery periods (at an RPE of 3–4). The total duration of the interval phase progressed from 20 minutes in the first two weeks to 30 minutes by week 6, and up to 45 minutes by the end of the intervention. Sessions concluded with a 10-minute cool-down. All HIIT sessions were monitored using Polar heart rate monitors, and intensity was prescribed relative to each participant’s maximum heart rate (HRmax). Progression was guided by weekly physiotherapist supervision and confirmed using the Borg RPE scale to ensure safe and effective workload adjustments. Participants were advised to exercise 3–5 times weekly and to recognise signs of hypoglycaemia and other exercise termination criteria, such as chest pain or shortness of breath. All supervised sessions were conducted in the late afternoon (between 3:00 and 7:00 pm), and participants were encouraged to maintain this schedule for their home-based sessions to ensure consistency across groups. This timing was selected to accommodate work schedules and because emerging evidence suggests that afternoon or evening exercise may enhance glycaemic control more effectively than morning exercise in individuals with T2DM [30–32].

The exercise programme targeted major muscle groups and was initially prescribed as 1–2 sets of 5 repetitions, with both the number of sets and repetitions progressively increased every two weeks according to participant tolerance. By the end of the 12-week programme, individual session durations ranged from 30 to 40 minutes. Adherence was recorded as the proportion of supervised sessions attended, with participants also maintaining logbooks and receiving weekly follow-up calls to encourage engagement (≥70% attendance was required). Compliance was defined as the proportion of attended sessions in which the prescribed training intensity (target heart rate and Borg RPE scale) was achieved. Attendance was documented by supervising physiotherapists, and compliance was verified through heart rate monitor recordings

The control group received standard hospital care without a structured exercise program. All the participants were evaluated at baseline and after 12 weeks. For illiterate patients, caregivers assisted them in understanding and adhering to the program. The exercise protocols were designed based on Maiya et al. (2006) [33] and Amaravadi [18] aligned with the American College of Sports Medicine (ACSM) guidelines [34]. Intervention details are summarized according to the FITT-VP framework [35] in Appendix 1 to improve clarity and reproducibility.

Outcomes and measurements

The primary outcome measures in this study included fasting insulin levels (FI) [mIU/L], Homeostatic Model Assessment of Insulin Resistance (HOMA-IR), and quality of life assessed using the World Health Organization Quality of Life-BREF (WHOQOL-BREF) questionnaire (Cronbach’s α = 0.73–0.81) [36]. The secondary outcome measures included functional aerobic capacity, assessed by using the six-minute walk test (6MWT) with the outcome expressed as distance walked in meters; body composition parameters, including subcutaneous fat, visceral fat, and muscle mass, were assessed using a bioelectrical impedance analyzer (Tanita MC-780MA, Tanita Corporation, Tokyo, Japan), fasting glucose levels (FG) [mg/dL], postprandial blood sugar levels (PPBS) [mg/dL], HbA1c levels [%], assessment of physical activity using the Global Physical Activity Questionnaire (GPAQ) (Cronbach’s α = 0.83–0.96) [37]. Waist and hip circumferences were measured using a flexible, non-stretchable tape measure following the standardised protocols of the International Society for the Advancement of Kinanthropometry (ISAK). Measurements were taken twice, and the mean value was used for analysis [38]. The three groups were assessed using these outcome measures at baseline and after the 12-week intervention period. All outcome assessments were conducted in the morning following an overnight fast of at least 10 hours. Participants were instructed to refrain from vigorous physical activity and caffeine consumption for at least 24 hours before testing to reduce potential confounding effects. All outcome assessors were blinded to group allocation throughout the study. The research staff responsible for conducting clinical, functional, and questionnaire-based assessments were not involved in the intervention delivery and had no access to participant randomization information, thereby minimizing the risk of detection bias.

Adverse events were monitored throughout the intervention by the principal investigator, who supervised all exercise sessions. Participants were instructed to report any injuries, hypoglycaemic episodes, or other untoward effects. All events were documented using structured reporting forms and reviewed weekly by the principal investigator. Adherence was tracked using participant logbooks and weekly follow-up calls.

Statistical analysis

Baseline characteristics were summarised as mean ± standard deviation (SD) for continuous variables and n (%) for categorical variables. Between-group comparisons were examined using one-way analysis of variance (ANOVA) for continuous variables and Chi-square (χ²) tests for categorical variables. Given the relatively small sample size and the large number of baseline variables, P-values were adjusted using both the Benjamini–Hochberg False Discovery Rate (FDR, q = 0.05) and Bonferroni correction (m = 23 tests; adjusted α = 0.002). These adjusted P-values are presented in Table 1 to address reviewer concerns regarding multiple comparisons. Consistent with CONSORT 2025 guidance [26], these results are descriptive and were not used to determine baseline equivalence; instead, baseline imbalances were subsequently included as covariates in the linear mixed-effects models.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Baseline Demographic and Clinical Characteristics of Participants (Adjusted for Multiple Comparisons).

https://doi.org/10.1371/journal.pone.0336898.t001

Statistical analyses were performed using the RCTapp Shiny interface [39]. The primary analysis was conducted using a linear mixed-effects model with participants specified as random effects and fixed factors for group (HIIT, A + R, Control), time (Pre, Post), and the group × time interaction. To account for observed baseline imbalances, models were covariate-adjusted for age, duration of type 2 diabetes mellitus (T2DM), baseline fasting glucose (FG), and baseline WHOQOL-BREF physical domain scores. This adjustment ensured that estimated treatment effects reflected true intervention-related changes rather than pre-existing group differences. Two-sided P-values < 0.050 were considered statistically significant. Between-group mean differences and corresponding 95% confidence intervals (CIs) from the covariate-adjusted model were reported for interpretability. Given the number of primary and secondary outcomes, the potential for Type I error inflation was acknowledged. Although P-values for inferential outcomes (Table 3) were not formally adjusted, interpretation emphasised effect sizes and confidence intervals over nominal significance. This approach aligns with PLOS ONE and CONSORT 2025 guidance, which prioritise estimation-based reporting over strict multiplicity correction in exploratory analyses.

Sample size calculation

Sample size estimation was conducted using G*Power version 3.1.9.7, based on a repeated-measures ANOVA (within-between interaction) to detect changes in clinical outcomes across three groups: high-intensity interval training (HIIT), combined aerobic and resistance training (A + R), and a control group receiving standard care. The calculation assumed a moderate effect size (f = 0.30), with an alpha level of 0.05, power of 0.80, two measurement time points (baseline and 12 weeks post-intervention), and an assumed correlation of 0.5 between repeated measures. These parameters indicated that a minimum of 66 participants (22 per group) would be required to detect significant group-by-time interaction. To account for an anticipated 20% dropout rate, the final target sample size was adjusted to 84 participants (28 per group). The assumed moderate effect size (f = 0.30) was based on prior trials of structured exercise interventions in T2DM, which reported mean effects ranging from f = 0.25–0.35 for HOMA-IR and HbA1c outcomes [18,40,41]. This value was therefore considered clinically meaningful and realistic for the anticipated metabolic improvements.

Characteristics of participants

Table 1 summarizes the baseline demographic, clinical, glycaemic, and functional characteristics of participants across the three groups: Combined Aerobic + Resistance (A + R), High-Interval Training (HIIT), and Control (n = 30 each). Values are presented as mean ± standard deviation (SD) for continuous variables and n (%) for categorical variables. Between-group comparisons were performed using one-way analysis of variance (ANOVA) for continuous variables and Chi-square (χ²) tests for categorical variables. To control for the increased risk of Type I error arising from the large number of comparisons relative to the modest sample size, P-values were adjusted using both the Benjamini–Hochberg False Discovery Rate (FDR, q = 0.05) and the Bonferroni correction (α = 0.002, m = 23 tests). After adjustment, only a subset of variables remained statistically significant under the FDR threshold, specifically the duration of type 2 diabetes mellitus (T2DM), fasting glucose (FG), post-prandial blood sugar (PPBS), muscle mass (MM), six-minute walk distance (6MWD), and WHOQOL-BREF physical, psychological, and environmental domains. These findings reflect true baseline imbalances, which were subsequently included as covariates in the mixed-effects model to ensure accurate estimation of intervention effects. Table 2 provides a summary of comorbidities and diabetes-related health behaviors.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Baseline Comorbidities and Health Behaviors of Participants.

https://doi.org/10.1371/journal.pone.0336898.t002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Changes in metabolic, body composition, and quality-of-life outcomes following 12 weeks of exercise intervention.

https://doi.org/10.1371/journal.pone.0336898.t003

HIIT: high-intensity interval training; OHA: oral hypoglycaemic agent. BMI: body mass index; T2DM: type 2 diabetes mellitus; HbA1c: glycated haemoglobin; FG: fasting glucose; PPBS: postprandial blood sugar; HOMA-IR: homeostatic model assessment of insulin resistance; 6MWT: six-minute walk test; GPAQ: Global Physical Activity Questionnaire.

Fasting glucose (FG): Compared with Control, HIIT reduced fasting glucose (MD −29.09 mg/dL, 95% CI −41.16 to −17.02) and A + R also improved glycaemic control (MD −20.59 mg/dL, 95% CI −30.96 to −10.21). The A + R vs. HIIT contrast was small and imprecise (MD 8.51 mg/dL, 95% CI −2.01 to 19.02).

HbA1c: Both interventions lowered HbA1c relative to Control: HIIT (MD −3.35%, 95% CI −4.11 to −2.58) and A + R (MD −3.33%, 95% CI −4.03 to −2.62). A + R vs. HIIT showed no meaningful difference (MD 0.02%, 95% CI −0.69 to 0.73).

Fasting insulin (FI) and insulin resistance (HOMA-IR): Fasting insulin decreased vs. Control in HIIT (MD −7.16 mIU/L, 95% CI −10.04 to −4.28) and A + R (MD −8.87 mIU/L, 95% CI −11.77 to −5.97). HOMA-IR improved in A + R relative to Control (MD −2.33, 95% CI −3.63 to −1.03), with a non-significant trend favoring HIIT (MD −1.17, 95% CI −2.47 to 0.13).

Post-prandial blood sugar (PPBS): Between-group estimates did not show a beneficial effect for HIIT vs. Control (MD 5.35 mg/dL, 95% CI −19.76 to 30.47) and indicated worse PPBS for A + R relative to Control (MD 36.11 mg/dL, 95% CI 7.97 to 64.26); A + R vs. HIIT favoured HIIT (MD −30.76 mg/dL, 95% CI −52.19 to −9.34). These findings should be interpreted in the context of broader glycaemic improvements (fasting glucose, HbA1c, fasting insulin) that consistently favored both active interventions.

Functional capacity (6-minute walk distance): Compared with Control, HIIT improved to 6MWD (MD + 178.91 m, 95% CI 130.47 to 227.35) and A + R yielded a larger gain (MD + 233.61 m, 95% CI 191.75 to 275.47). A + R exceeded HIIT (MD + 54.71 m, 95% CI 14.92 to 94.50).

Body composition: We observed favourable between-group effects for both interventions versus Control. Fat-free mass increased with HIIT (MD + 7.54 kg, 95% CI 4.71 to 10.36) and A + R (MD + 5.96 kg, 95% CI 3.06 to 8.86); subcutaneous fat (HIIT MD −7.16%, 95% CI −9.33 to −4.99; A + R MD −8.37%, 95% CI −10.65 to −6.09) and visceral fat (HIIT MD −4.70%, 95% CI −5.93 to −3.47; A + R MD −4.58%, 95% CI −5.86 to −3.31) decreased. The A + R–HIIT contrasts were small with CIs crossing zero, indicating no clear difference between the active interventions.

Physical activity and sedentary time: Compared with Control, total physical activity (GPAQ MET-min/week) increased for HIIT (MD + 845.29, 95% CI 292.06 to 1398.51) and A + R (MD + 1137.51, 95% CI 577.09 to 1697.93). Sedentary time decreased for HIIT (MD −653.42 min/week, 95% CI −1087.96 to −218.87) and A + R (MD −428.81 min/week, 95% CI −849.88 to −7.73). The A + R vs HIIT contrasts for these behavioral outcomes were imprecise.

Quality of life (WHOQOL-BREF): All four domains improved versus Control in both interventions. Representative effects include the physical domain (HIIT MD + 10.29, 95% CI 4.06 to 16.51; A + R MD + 13.77, 95% CI 6.62 to 20.91), psychological (HIIT MD + 17.26, 95% CI 10.55 to 23.97; A + R MD + 27.80, 95% CI 20.64 to 34.96), social (HIIT MD + 23.93, 95% CI 16.53 to 31.34; A + R MD + 32.20, 95% CI 24.53 to 39.87), and environmental (HIIT MD + 39.60, 95% CI 33.51 to 45.69; A + R MD + 43.37, 95% CI 35.62 to 51.13).

Adverse events

No adverse events were reported by participants in either intervention group. Adherence to the exercise programme was defined as completion of >70% of prescribed sessions, and most participants met this criterion.

Blood glucose regulation and insulin sensitivity

Both exercise groups showed significant within-group improvements from baseline in key glycaemic outcomes, including fasting blood sugar, postprandial blood sugar, fasting insulin, and HOMA-IR. Between-group analyses indicated that these improvements were greater in the A + R and HIIT groups than in the control group, confirming the efficacy of structured exercise for metabolic control in T2DM. Direct comparisons between A + R and HIIT did not identify statistically significant differences on these primary glycaemic outcomes, indicating no clear superiority of one modality over the other. Magnitudes and precision of effects are reported as Cohen’s d with 95% confidence intervals in the Results and in Table 3.

For fasting and postprandial blood sugar, HIIT exhibited numerically larger within-group reductions, consistent with mechanisms such as enhanced mitochondrial biogenesis, increased GLUT4 translocation, and improved skeletal muscle glucose uptake during repeated high-intensity efforts [42,43]. These observations align with prior reports that interval-based training can elicit rapid glycaemic benefits in time-efficient formats. However, the differences relative to A + R did not reach statistical significance, underscoring that both interventions are comparably effective for glucose regulation.

Both interventions also reduced fasting insulin and improved HOMA-IR compared with baseline and control, supporting enhanced peripheral insulin sensitivity with structured training. Although numerically greater declines in fasting insulin were observed with HIIT, between-group differences were not statistically significant, indicating that either modality can be used to improve insulin dynamics depending on patient preference and feasibility.

HbA1c also decreased significantly in both intervention arms, confirming improvements in long-term glycaemic regulation. A + R produced slightly greater absolute reductions, a profile consistent with the prolonged cardiovascular stimulus and additive resistance component characteristic of combined training. These findings echo evidence that both interval and continuous training lower HbA1c through distinct physiological pathways [44]. The between-group difference, however, was not statistically significant, supporting therapeutic equivalence of A + R and HIIT for this outcome.

Our pattern of effects aligns with recent evidence. The apparent advantage of HIIT for fasting glucose reduction is concordant with meta-analytic findings that higher-intensity interval formats enhance glucose regulation and insulin resistance [44], while the broader benefits observed with combined aerobic and resistance training on HbA1c, adiposity indices, and quality of life echo reports that multi-component programmes confer wider cardiometabolic and psychosocial gains [21,41]. Narrative and trial data published in 2024–2025 extend this view: Al-mhanna et al. highlighted that exercise effects on lipid metabolism in overweight or obese adults with T2DM are sensitive to programme design and participant heterogeneity [22], and Tayebi et al. demonstrated clinically relevant improvements after a 12-week HIIT protocol in obese men, while calling for inclusion of women and older adults to strengthen generalisability [23]. These converging findings support the external validity of our results in diverse service contexts. To aid clinical interpretation, we report effect sizes with 95% confidence intervals for all primary outcomes in the Results and in Table 3, which clarifies the magnitude and precision of treatment effects.

Quality of life and functional outcomes

Improvements in functional capacity were reflected in significant increases in six-minute walk distance (6MWD) in both exercise groups, underscoring the role of structured activity in enhancing cardiorespiratory fitness and endurance in T2DM patients. A + R elicited greater gains, due to its continuous aerobic stimulus combined with resistance training, which improves muscular efficiency and exercise tolerance. The effect of functional enhancement is especially critical in this population given the link between physical inactivity, mobility limitations, and disease progression.

In terms of quality of life, both interventions produced significant gains across all four WHOQOL-BREF domains. A + R demonstrated superior improvements in physical, psychological, and social domains, suggesting it is particularly effective in promoting holistic well-being. HIIT, on the other hand, showed the greatest benefit in the environmental domain, potentially reflecting improved perceived autonomy, accessibility, and adaptability to lifestyle contexts. These domain-specific gains are congruent with the findings of Collins et al. (2021), who reported enhanced psychosocial outcomes following structured aerobic routines in individuals with chronic illnesses [9]. The social benefits observed with A + R may also stem from its moderate intensity, which may facilitate greater adherence and enjoyment, thereby positively reinforcing self-efficacy and patient motivation.

Our results mirror those of Gajanand et al. (2023), who demonstrated that low-volume combined aerobic and resistance HIIT improved not only physiological markers but also health-related QoL metrics among people living with T2DM [19]. This reinforces the potential for blended training protocols to deliver comprehensive health improvements across both physiological and psychosocial domains.

Body composition adaptations

Both exercise interventions improved body composition, although the nature of changes differed. A + R produced greater reductions in subcutaneous and visceral fat, which is consistent with its role in enhancing fat oxidation and lipid metabolism, outcomes that are particularly important given the established links between visceral adiposity, systemic inflammation, and insulin resistance in T2DM. By contrast, participants in the HIIT group demonstrated measurable increases in lean mass, suggesting potential anabolic adaptations through the recruitment of type II muscle fibres and stimulation of protein synthesis pathways during repeated high-intensity bouts. Such adaptations may be clinically relevant for older adults or those with sarcopenic obesity, where the preservation of muscle mass is vital for functional independence and metabolic resilience.

These findings should, however, be interpreted with caution. Increases in lean mass of this magnitude are uncommon in this population over short intervention periods, and the use of bioelectrical impedance analysis (BIA) may have contributed measurement variability, as results can be influenced by hydration status and other physiological factors. While our observations are directionally consistent with previous reports showing modest hypertrophic effects of high-intensity exercise [45,46], the potential influence of methodological limitations cannot be excluded. Taken together, these results suggest complementary benefits of both modalities: A + R is more effective for reducing adiposity, while HIIT may contribute to lean mass improvements. Future studies employing more robust assessment methods such as DXA or MRI are warranted to validate these outcomes.

Implications for exercise prescription

The current findings reinforce the importance of adopting a tailored approach to exercise prescription for individuals with T2DM. Both HIIT and A + R demonstrated clinically significant benefits, but their respective advantages suggest that different patient profiles may be better suited to one modality over the other. Importantly, this study affirms that structured physical activity when appropriately matched to individual needs and capacities can offer multifaceted improvements in metabolic, physical, and psychosocial outcomes.

HIIT is ideally suited for patients who require or prefer time-efficient interventions to address glycaemic dysregulation and insulin resistance. Its efficacy in rapidly improving FG and HOMA-IR, alongside its capacity to promote muscle hypertrophy, positions HIIT as a powerful tool in metabolic rehabilitation particularly among younger adults, those with limited time availability, or patients prioritising improvements in insulin sensitivity and lean body mass. However, HIIT’s demanding nature may pose adherence challenges for certain populations, including older adults or those with comorbidities.

Combined Aerobic and Resistance Training (A + R) is recommended for individuals seeking more holistic, gradual improvements in cardiovascular health, glycaemic control, body composition, and quality of life. The moderate intensity of this intervention enhances long-term adherence and appears to offer broader psychosocial benefits, particularly in the domains of psychological and social well-being. A + R may be more suitable for patients at earlier stages of diabetes progression or those with physical or psychological barriers to engaging in high-intensity protocols.

Clinicians should consider not only physiological outcomes but also patient preferences, accessibility, perceived enjoyment, and risk profiles when recommending exercise regimens. These findings lend further support to the development of personalised, patient-centred strategies in diabetes rehabilitation.

Study strengths and limitations

This study offers robust evidence for the comparative efficacy of two distinct exercise modalities within a randomised controlled design. The integration of both subjective and objective outcomes enhances the ecological validity of the findings, while the use of a rigorous statistical analysis adds methodological rigour.

However, certain limitations must be acknowledged: The findings are based on participants who adhered to their prescribed regimens, which may limit generalisability to patients with advanced diabetic complications. The study duration was limited to 12 weeks; longer-term follow-up is necessary to assess sustainability and the durability of observed effects on body composition and glycaemic control. Real-world implementation may vary, particularly with HIIT, which requires higher effort and supervision that may not always be feasible in unsupervised or community settings.

Directions for future research

Future investigations should aim to:

Explore combination regimens: Assessing whether integrating HIIT with A + R in phased or concurrent formats leads to additive or synergistic effects.

Evaluate long-term outcomes: Trials extending beyond six months will be critical to understanding chronic adaptations in glycaemic control, cardiovascular fitness, and patient adherence.

Conduct subgroup analyses: Examining the differential responses in various T2DM populations, such as older adults, individuals with multiple comorbidities, or those at high cardiovascular risk, will offer deeper insights into personalised exercise planning.

These research directions will further inform clinical guidelines and enhance the translation of evidence-based exercise strategies into sustainable diabetes care models.