figure a

Introduction

Alterations in arterial structure and stiffness are major culprits for the increased risk of cardiovascular disease (CVD) and early mortality in individuals with type 2 diabetes [1, 2]. Pulse wave velocity (PWV) and carotid intima–media thickness (IMT) are established surrogate measures of arterial stiffness and structure, respectively, that are strong predictors of cardiovascular outcomes [2,3,4]. Hence, interventions in individuals with type 2 diabetes that can improve these markers of arterial health are critically important for preventing and/or delaying adverse cardiovascular events and mortality. Exercise training of different types and intensities has been shown to be an effective strategy for improving PWV and IMT in individuals with type 2 diabetes [5,6,7,8].

Many of the cardiovascular benefits of exercise are associated with improvements in cardiorespiratory fitness (CRF), as determined by \( \dot{V}{\mathrm{O}}_{2\mathrm{peak}} \), independent of traditional risk factors [9]. A higher CRF is not only associated with lower levels of risk factors for the metabolic syndrome, lower body fat, improved insulin resistance and BP, and reduced inflammation [10], all of which are independently associated with arterial structure and stiffness [11], but it is also independently associated with reduced carotid IMT [12] and increased PWV in individuals with and/or at risk for type 2 diabetes [13]. However, there is still wide interindividual variability in the response to exercise, with certain individuals improving CRF with exercise training, while others do not [14, 15]. In fact, the theme of CRF response to exercise was recently highlighted in a consensus statement by an expert panel; the authors recommended future research to better understand the interindividual variability in the therapeutic effect of exercise [15].

Although CRF, based on \( \dot{V}{\mathrm{O}}_2 \), is connected with many health outcomes [10], previous investigations have suggested that exercise and being physically active can have beneficial effects on health that are independent of improvements in CRF, in both healthy individuals [16], and those with increased cardiovascular risk [17, 18]. However, in individuals with type 2 diabetes, there has only been one investigation that has addressed the impact of CRF response on health outcomes (the Health Benefits of Aerobic and Resistance Training in individuals with Type 2 Diabetes [HART-D] study), in which improvements in glycaemic control and body fat were observed regardless of improvements in CRF following a 9 month exercise intervention [14]. It is particularly important to consider the possibility of having health benefits from exercise despite alterations in CRF when working with individuals with type 2 diabetes as they are more likely not to experience improvements in CRF following exercise due to having slowed oxygen uptake, impaired muscle oxygen delivery, premature muscle fatigue and reduced overall exercise tolerance as a consequence of the disease [19]. No investigation, however, has assessed the long-term relationship between CRF non-response and measures of vascular stiffness and structure in individuals with type 2 diabetes. Given that adverse vascular changes are thought to be the major mechanism connecting diabetes with increased risk of CVD morbidity and mortality [2], and that functional and structural arterial properties are highly related to CRF [12, 13], it is uncertain whether exercise-induced vascular changes only occur in those who also have improved CRF. Therefore, we investigated arterial stiffness and structure indices in individuals with type 2 diabetes who either improved or did not improve CRF following a 1 year exercise intervention. We hypothesised that, following 1 year of exercise training, participants with type 2 diabetes would have improvements in arterial stiffness and structure irrespective of improvements in CRF, as compared with a non-exercise control group.

Methods

Study design and participants

The current investigation assessed individuals with type 2 diabetes who took part in the D2FIT study. The design and methodology of the D2FIT study is fully described elsewhere [20]. In brief, the D2FIT study was a 1 year exercise randomised controlled trial with three distinct arms: a non-exercise control group, a moderate continuous training (MCT) with resistance training (RT) group, and a high-intensity interval training (HIIT) with RT group (ClinicalTrials.gov registration no. NCT03144505). The three groups were allocated with a 1:1:1 ratio by a researcher external to the project, using a computer-generated list of random numbers. The primary outcome of this study concerned changes in HbA1c following exercise in individuals with type 2 diabetes, with secondary outcomes focused on several health-related parameters, including CRF and vascular health changes, which were assessed at baseline and at follow-up (1 year). Inclusion criteria consisted of the following: adults previously diagnosed with type 2 diabetes in accordance with ADA criteria [21]; aged between 30–75 years old; no major microvascular complications from diabetes, such as nephropathy, neuropathy and retinopathy, when considered severe enough to be contraindicative to exercise, as assessed by a clinical physician; no macrovascular complications, such as coronary artery disease and cerebral vascular disease; BMI <48 kg/m2; and no physical limitations that would prevent individuals from practicing exercise. Participants were recruited within the Lisbon Metropolitan Area (Portugal) between February 2014 and July 2016. Power and sample size calculations (G-Power, Version 3.1.3; Düsseldorf, Germany) were based on a predicted change in HbA1c of 0.66 units with an SD of 1.2 units, α = 0.05, 1-β = 0.80, and a 10% expected dropout rate. For the D2FIT project, a total of 80 individuals were selected and randomised using a computer-generated list of random numbers. However, for this specific analysis, we considered only those who completed the investigation, resulting in a final sample size of 63 participants with type 2 diabetes.

Written informed consent was obtained from all participants prior to screening. The study protocol was reviewed and approved by the Ethics Committee of the Portuguese Diabetes Association (approval number: 07/17/2013).

Exercise intervention

Exercise prescription for both exercise groups was standardised, according to body weight and based on physical activity guidelines [22], to achieve a weekly target of 41.84 kJ/kg (10 kcal/kg) for both exercise groups.

Participants in the HIIT with RT (n = 19) and MCT with RT groups (n = 22) exercised 3 days per week. Throughout the whole year of the intervention, individuals in both groups participated in a planned periodisation exercise prescription programme with individualised intensity based upon heart rate reserve (HRR), to gradually introduce exercise prescription. A heart rate Polar band (Polar T-31; Bethpage, NY, USA) was worn by participants during the exercise sessions and was used to monitor the intensity of training. The MCT group performed continuous cycling (Monark Ergometric 828e (Vansbro, Sweden) at 40–60% of HRR throughout the 1 year intervention. The HIIT group performed 1 min of cycling at 90% of HRR, followed by a 1 min rest period at 40–60% of HRR, using a 1:1 exercise:rest ratio. Following the aerobic exercise component, both groups underwent whole-body RT, which included one set of 10–12 repetitions of upper- and lower-limb exercises. All exercise sessions were overseen and supervised by a certified exercise physiologist. A full detailed description of the periodisation protocol can be found elsewhere [20].

A standard counselling session regarding general physical activity guidelines was provided for the control group (n = 22). Additionally, the control group, along with both exercise groups, met once every 4 weeks for thematic sessions, with topics including nutrition or physical activity, to improve trial adherence.

Anthropometry

Participants were weighed without shoes and wearing minimal clothing, on an electronic scale (Seca, Hamburg, Germany), to the nearest 0.01 kg. Height was measured to the nearest 0.1 cm with a stadiometer (Seca), according to standardised procedures [23]. BMI (kg/m2) was calculated as the ratio between weight (kg) and the square of height (m).

Brachial BP

After 15 min in the supine position, the right and left brachial systolic BP (SBP) and diastolic BP (DBP) were measured using an automated oscillometric cuff (HEM-907-E; Omron, Tokyo, Japan). Two measurements were taken, with a third measurement being performed if duplicate values deviated by >5 mmHg. The mean arterial pressure (MAP) was calculated using the formula: MAP = DBP + [1/3(SBP − DBP)]. Pulse pressure was calculated as follows: pulse pressure = SBP − DBP.

Carotid artery IMT

Carotid artery IMT was defined as the distance between the leading edge of the lumen–intima interface and the leading edge of the media–adventitia interface of the far wall of the right carotid artery. This measurement was performed using an ultrasound scanner equipped with a linear 13 MHz probe (MyLab One, Esaote, Italy) [24] and distension curves were acquired within a segment of the carotid artery about 1 cm before the flow divider.

Carotid arterial stiffness indices

After 15 min in a supine position, the carotid arterial stiffness measurement was conducted on the right side of the body. An ultrasound scanner equipped with a linear 13 MHz probe (MyLab One), with Quality Arterial Stiffness technology, was used approximately 1 cm before the bifurcation. This allows the calculation of carotid stiffness indices: PWV (m/s), carotid distensibility coefficient (1/KPa) and stiffness index β. A detailed description of the calculations for the carotid stiffness indices can be found in the electronic supplementary material (ESM) Methods.

Contralateral PWV

PWV was measured by applanation tonometry immediately after ultrasound imaging. A single operator located the carotid, femoral, radial and distal posterior tibial arteries on the left side of the body and marked the point for capturing the corresponding pressure curves with two specific pressure sensitive transducers. The distance between the carotid and femoral, and radial and distal posterior tibial arteries were measured and values were directly inserted into the Complior Analyse software (ALAM Medical, Paris, France). Detailed description of the methods for contralateral PWV are available in the ESM Methods. PWV values obtained from measurements of the carotid to femoral artery, carotid to radial artery, and carotid to distal posterior tibial artery were taken as indices of aortic and peripheral arterial stiffness for upper and lower limbs, respectively. The quality of the PWV records was immediately evaluated by a second observer. Whenever a continuous decrease before the sharp systolic upstroke was not clearly seen, or the tolerance was above 5 m/s, a second measure was taken.

Carotid BP

Carotid SBP was obtained from left carotid traces acquired during the PWV assessment, from the Complior Analyser (ALAM Medical). The waveforms were averaged and the mean values were extracted from 15 s window of acquisition. The carotid waveforms were calibrated with left brachial MAP measured immediately before the acquisition.

CRF

For the assessment of CRF, a Bruce standard protocol [25] was used on a motorised treadmill (Quinton, model Q-65; Cardiac Science, Bothell, WA, USA) to exhaustion, in a spacious ventilated room, with only the participant, the technician and the cardiologist present during the test. All graded exercise tests were monitored using a 12-lead ECG PC-based acquisition module (model Quark C12; Cosmed, Rome, Italy) and all data, including heart rate, were monitored and recorded using Cosmed software. Expired gases were continuously analysed, breath-by-breath, through a portable gas analyser (K4b2; Cosmed). To ensure peak effort, participants attained volitional fatigue and met at least one of the following objective test criteria: (1) a respiratory exchange ratio of 1.1 or higher; (2) participants reached predicted maximal heart rate; (3) oxygen uptake did not increase in spite of increasing workload. However, if there was a medical contraindication, the exercise testing was interrupted. The highest 20 s value for \( \dot{V}{\mathrm{O}}_{2\mathrm{peak}} \) (l/min) attained in the last min was used in the analysis and is termed ‘CRF’ from here on.

Defining CRF improvement

CRF improvement was defined as having a change in \( \dot{V}{\mathrm{O}}_{2\mathrm{peak}} \) (l/min) ≥5.0% from baseline to follow-up. This cut-off point was chosen based on the technical error of measurement (TEM) of \( \dot{V}{\mathrm{O}}_{2\mathrm{peak}} \) using the Cosmed K4, which is ~4.5% [26]; this was also the cut-off point used by Pandey et al in a similar analysis with a type 2 diabetes population [14]. Any value greater than the TEM is more likely to be a true biological response to exercise training rather than a result of the typical error variation present in the measurement technique and the day-to-day biological variability present in the individual [27].

Laboratory measurements

Participants underwent biochemical assessments, including the analysis of HbA1c. Blood samples were collected from an indwelling catheter and drawn into chilled, heparinised tubes and centrifuged rapidly to avoid glycolysis. HbA1c was analysed by immunoassay (Premier Hb9210 HbA1c Analyzer; Trinity Biotech, Wicklow, Ireland).

Physical activity

Physical activity was assessed using an accelerometer (ActiGraph, GT3X+ model; Fort Walton Beach, FL, USA). Participants were asked to wear the accelerometer on the right hip, close to the iliac crest, for 7 consecutive days. The device activation, download and processing were performed using the software, Actilife (v.6.9.1; ActiGraph). The devices were activated on the first day, in the morning, and data were recorded using the raw mode with a 100 Hz frequency, and posteriorly downloaded into 15 s epochs. Apart from actual non-wear time (i.e. when the device was removed during sleep and water activities), periods of at least 60 consecutive min of zero activity intensity counts were considered as non-wear time. A valid day was considered as having 600 min or more of wear time. Participants with at least 4 valid days (including 1 weekend day) were included in the analyses. All minutes for which the accelerometer counts were below 100 cpm were defined as sedentary time. Accelerometer counts ≥100 cpm and <2020 cpm were classified as low-intensity physical activity (LIPA). Accelerometer counts ≥2020 cpm were classified as moderate-to-vigorous physical activity (MVPA) [28].

Statistical analysis

Descriptive statistics, including measures of central tendency (mean) and variability (SD) for normally distributed variables, and median (interquartile range) for skewed variables, were used to describe the baseline descriptive characteristics of the control group and those in the exercise groups who had improved (CRF responders) and did not have improved (CRF non-responders) CRF. To compare differences in descriptive characteristics and changes in physical activity between the three groups, a one-way ANOVA with Bonferroni adjustment for multiple comparisons or the non-parametric Kruskal–Wallis test was used to compare differences in continuous measures, and a χ2 test or Fisher’s exact test was used for categorical variables.

Generalised estimating equations were used to model outcomes using sex and total number of training sessions completed as covariates. Since arterial stiffness is dependent on arterial pressure [29], per cent change in MAP from baseline to the end of the study was included as a covariate in all regression models assessing differences in arterial stiffness and structure indices. Further adjustment for baseline MVPA and the exercise intervention group (i.e. non-exercise control, HIIT or MCT) were also performed given the significant difference between CRF responder groups at baseline. A least significant difference post hoc test was used to estimate the between- and within-group effects with regard to haemodynamic and vascular outcomes. A linear distribution for the response was assumed and an autoregressive correlation matrix was set to the data.

A p value of <0.05 was considered statistically significant. Data analyses were performed using IBM SPSS Statistics version 22.0 (SPSS, Chicago, IL, USA) and STATA version 13.1 (StataCorp, College Station, TX, USA).

Results

Of the 41 individuals with type 2 diabetes who engaged in 1 year of either MCT or HIIT combined with RT, 14 had clinically meaningful improvements in CRF (mean \( \Delta \dot{V}{\mathrm{O}}_{2\mathrm{peak}} \), 0.24 l/min), whereas 27 participants did not have improvements in CRF (mean \( \Delta \dot{V}{\mathrm{O}}_{2\mathrm{peak}} \), −0.16 l/min) following exercise training. There were no significant difference in the number of training sessions that were attended by those in whom CRF was improved and not improved (p = 0.12). Baseline descriptive characteristics of the participants are presented in Table 1. At baseline there were no differences (p>0.05) between the control group, those in whom CRF was improved and those in whom CRF was not improved, with the exception of MVPA, which was higher (p<0.05) in CRF non-responders compared with the control group (p<0.05), and the exercise intervention group category, with more of the CRF non-responders being in the HIIT group compared with CRF responders (p<0.05).

Table 1 Baseline descriptive characteristics of individuals in the control group, and exercise groups with either improved or unimproved CRF following 1 year of exercise training
Full size table

Time in LIPA and MVPA were not significantly different between the CRF responders (ΔLIPA, 6.9 ± 57.9 min; ΔMVPA, 1.1 ± 18.0 min), non-responders (ΔLIPA, −11.1 ± 40.8 min; ΔMVPA, −2.8 ± 22.9 min) and control participants (ΔLIPA, 0.2 ± 70.5 min; ΔMVPA, −0.1 ± 12.6 min).

Table 2 presents the between- and within-group changes for haemodynamic measures, and local and regional indices of early pathological changes in arterial structure and stiffness, after adjusting for sex, total number of training sessions completed, and exercise intervention group. Arterial structure and stiffness indices were additionally adjusted for per cent change in MAP. Regardless of whether CRF improved, those in the exercise groups had either decreased (CRF non-responders: −0.03 mm) or unchanged (CRF responders: 0.001 mm) carotid IMT compared with the control group (+0.04 mm), following the 1 year intervention. Improvements (p<0.05) were also observed in peripheral arterial stiffness in exercisers, regardless of improvements in CRF, as lower-limb PWV was decreased vs the control group (CRF responders: −0.38 m/s; CRF non-responders: −0.40 m/s; control: +1.31 m/s). However, only the CRF responders had significantly decreased PWV of the upper limb compared with control participants (CRF responders: −1.29 m/s; control: +0.12 m/s). As for central stiffness, only the CRF non-responders had significant changes in carotid distensibility coefficient (p<0.05) compared with the control group. Despite changes in aortic PWV not differing from the control group (p>0.05), changes in aortic PWV differed (p<0.05) between CRF responders and non-responders, with non-responders having increased aortic PWV (CRF non-responders: 1.62 m/s), while the responders had decreased levels (CRF responders: −0.68 m/s) following 1 year of exercise training. No other interaction effects between individuals in the exercise groups who had improved or who did not have improved CRF vs individuals in the control group were found for the remaining vascular stiffness or haemodynamic measures (p>0.05). Further adjustment for baseline MVPA did not alter the results (data not shown).

Table 2 Within- and between-group changes in vascular function and stiffness indices for individuals in the control group, and exercise groups with either improved or unimproved CRF following 1 year of exercise training
Full size table

Figure 1 displays the individual per cent changes in local and regional indices of early pathological changes in indices of arterial structure and stiffness in the control participants and individuals in the exercise groups who either had increased or not improved post-trial CRF. Even among the participants in the exercise groups with and without improvements in CRF, there was a wide range of interindividual variability in vascular adaptations to exercise training. Overall, compared with the control group, those in the exercise groups had improvements (p<0.05) in lower-limb PWV and carotid IMT (Fig. 1c, d), regardless of whether they had concurrent improvements in CRF.

Fig. 1
figure 1

Individual per cent changes in (a) aortic PWV, (b) upper-limb PWV, (c) lower-limb PWV, (d) carotid IMT, (e) carotid PWV and (f) carotid distensibility coefficient (DC). *p<0.05 vs control after adjustment for sex, total number of training sessions completed and per cent change in MAP

Full size image

Discussion

The aim of this study was to assess whether improvements in vascular function were independent of improvements in CRF, based on \( \Delta \dot{V}{\mathrm{O}}_{2\mathrm{peak}} \), following 1 year of exercise training in individuals with type 2 diabetes. This is the first study to show that a long-term exercise intervention significantly decreased carotid IMT and lower-limb arterial stiffness in participants with type 2 diabetes, regardless of improvements in CRF. These findings are important given that vascular complications are common in individuals with type 2 diabetes and that these vascular alterations are a main mechanism connecting diabetes to increased risk of cardiovascular morbidity and mortality [1, 30]. Hence, our findings suggest that exercise-induced benefits in terms of vascular health may disrupt the close link between diabetes and CVD. In addition, individuals with type 2 diabetes are known to have CRF impairments independent of physical activity levels, obesity and presence of CVD [19], making them more likely to not experience changes in CRF following exercise, as compared with adults without type 2 diabetes [14, 31]. Thus, although exercise training-induced improvements in CRF have been correlated with decreases in CVD risk factors [10], this is the first study to show that the absence of an increase in CRF after exercise training does not preclude the possibility of beneficial vascular adaptations in individuals with type 2 diabetes.

In the present study, an interaction between CRF responders and non-responders was found in aortic PWV, such that individuals who had improved CRF with exercise had more favourable changes in aortic PWV. Strong associations between improvements in aortic vascular stiffness and reduced risk of cardiovascular and all-cause mortality have been previously reported [32]. However, using a central surrogate of PWV measured through carotid PWV, no differences between the exercise and control groups were found. Although carotid and aortic PWV are both classified as elastic and valid surrogate measures of central stiffness [2, 3], the ultrastructure of the carotid artery is more like the abdominal aorta than the ascending aorta. In addition, discrepancies between changes in carotid and aortic artery PWV following an intervention could also result from differences in measurement techniques: aortic PWV is determined manually by measuring the distance between the carotid and femoral sites, whereas carotid PWV is determined cross-sectionally from local changes in pressure and artery diameter due to the short vessel path (10–15 cm) and subsequent short time lag (5–15 ms) of the carotid artery, making manual measurement difficult [33]. Nevertheless, compared with the control group, decreases in PWV of the lower limb of exercisers with and without improvements in CRF were observed suggesting that exercise training may have a larger effect on peripheral rather than central stiffness, which may be attributable to the mechanical compression of the arteries due to muscular contraction during exercise [34]. In line with our findings, a recent meta-analysis of the impact of exercise on both central and peripheral arterial stiffness in adults, observed that exercise training had less of an effect on central arterial stiffness compared with peripheral measures [35]. Even though aortic PWV is the only segment of the arteries associated with decreased CVD risk, changes in lower-limb PWV following exercise might induce local adaptations and have implications on other systems, such as glucose transport and insulin signalling, in the muscle [36].

Beyond PWV, the distensibility coefficient, measured through ultrasound of the carotid artery, is also an indicator of vascular stiffness, which tends to deteriorate with ageing and type 2 diabetes [37, 38]. We found that, regardless of CRF improvement, 1 year of exercise training was able to halt the natural age- and disease-related deterioration in carotid distensibility vs baseline, such that decreases in carotid distensibility coefficient were only observed in control participants.

Exercise training has also been shown to improve indices of local vascular structure [6]. In accordance with previous findings, we also observed favourable effects of exercise training on carotid IMT independent of improvements in CRF. Given that for every 0.1 mm increase in carotid IMT, the risk of having a myocardial infarction increases by ~10% [4], our finding of a decrease in carotid IMT by 0.03 mm in the CRF non-responders and maintenance of carotid IMT in responders has important clinical implications in terms of decreasing cardiovascular events, which can occur regardless of having changes in CRF. In addition, it is important to note that carotid IMT is strongly linked to age and has been estimated to increase by ~0.010 mm per year in men and women [39]. Hence, our finding of a decrease and/or maintenance of carotid IMT after 1 year is even more relevant as it indicates that regardless of CRF improvements, long-term exercise can slow down the age-related increases in IMT, or even promote a decrease in carotid IMT.

Currently, we have only fragments of insight regarding the mechanisms involved in vascular remodelling following long-term exercise training regimes. Teasing apart the contribution of CRF on the vascular system [40] may provide key information for tailored interventions aiming for the prevention of arterial stiffening and related cardiovascular complications. In fact, most of the vascular benefits observed with exercise are thought to be mediated by the increased exercise-induced blood flow, which is responsible for increasing endothelial shear stress [11]. The force exerted by the blood flow on the arterial wall (i.e. shear stress) stimulates the production of nitric oxide, a potent vasodilator that adjusts the shear stress to which the artery is subjected to and plays a key role in repairing damaged endothelial cells, leading to an overall improvement in vascular health with medium-to-long-term implications on vascular structure and function [11]. This shear stress and pulsatile stretching of the arteries is also believed to aid in the breakdown of diabetes-associated collagen cross-links, resulting in decreased arterial stiffness [11]. Exercise-related improvements in vascular health may also be influenced by exercise-induced changes in metabolic disease risk factors, including hypertension, dyslipidaemia, inflammation and insulin resistance [41]. However, although CRF is associated with a favourable metabolic profile [40], several reports suggest that the exercise-induced metabolic improvements, which in turn improve vascular health, are independent of changes in CRF [18]. For instance, improvements in metabolic parameters, such as HbA1c and measures of adiposity, have been observed in individuals with type 2 diabetes, irrespective of improvements in CRF, following a 9 month exercise intervention [14]. Additionally, it has been suggested that the high interindividual variability in response to exercise can be partially due to alterations in non-exercise physical activity (NEPA), which is known to be affected by the intensity and duration of exercise training sessions [42]. However, we did not observe any differences in LIPA and MVPA levels between control participants and those in the exercise intervention, regardless of improvements in CRF, suggesting that improvements in vascular outcomes in the present study were not a consequence of higher NEPA.

Understanding the factors that lead to an increase in CRF in some but not all individuals undertaking an exercise intervention remains a source of considerable discussion [15]. Currently, there is no consensus on how to best quantify individual variability and categorise responders and non-responders to an exercise intervention. For instance, previous investigations have used various approaches to classify a CRF non-responder, including: a ≤0 l/min change in CRF [43, 44]; a non-clinically significant <5% change in CRF [14]; a within-participant CV of 5.6% from the literature [45]; arbitrary cut-off points of CRF ≤0 ml kg−1 min−1, ≤1.75 ml kg−1 min−1 and ≤3.5 ml kg−1 min−1 [46]; and a change in CRF ≤TEM [31, 47]. Nonetheless, all of these investigations have focused on the use of \( \dot{V}{\mathrm{O}}_2 \) (i.e. either maximal or peak) as the indicator of CRF. However, there are other indices of CRF that may provide additional information on the effectiveness of exercise interventions. Our results show that despite the absence of CRF improvements based on \( \dot{V}{\mathrm{O}}_{2\mathrm{peak}} \) in the non-responder group, both CRF responders and non-responders had increased time to exhaustion during maximal CRF exercise testing at 1 year vs baseline. It is possible that the increases in the time to exhaustion observed in the non-responder group were due to improvements in walking performance and other efficiency factors as a result of the exercise intervention, which may potentially contribute to more favourable effects on the economy of motion without leading to a concurrent increase in the \( \dot{V}{\mathrm{O}}_{2\mathrm{peak}} \) [48]. Moreover, previous investigations have suggested that there is substantial variability in the time to exercise exhaustion, with increases in time to exhaustion being independent of changes in \( \dot{V}{\mathrm{O}}_{2\max } \) over a series of repeated exercise tests [49, 50].

This study is not without limitations. Although our sample size was relatively large considering the length of the exercise intervention, the small number of participants within each group (i.e. control, CRF non-responders and CRF responders) could have affected the power to detect significant changes between groups.

In conclusion, 1 year of exercise training was able to slow the progression of the natural age- and disease-related deterioration of vascular health in individuals with type 2 diabetes, with significant improvements being observed in carotid IMT and lower limb arterial stiffness vs a non-exercise control group, regardless of changes in CRF. Thus, a lack of improvement in CRF following exercise in individuals with type 2 diabetes does not necessarily entail a lack of improvement in vascular health. From a clinical standpoint, when assessing the successfulness of an exercise intervention in individuals with type 2 diabetes, practitioners should not overlook the impact of exercise training on other health indicators by simply focusing on CRF, as other favourable health benefits are still obtainable with exercise in this patient population despite lack of improvements in fitness.