Exercise training improves cardiorespiratory fitness and oxygen uptake kinetics to a similar extent in early and late middle‐aged adults with type 2 diabetes Authors: Catherine Kiely, Eamonn O'Connor, Donal O'Shea, Simon Green, Mikel Egaña Published: 2026 ABSTRACT AbstractTo determine the exercise training–induced changes in peak oxygen uptake (V̇O2peak) and oxygen uptake kinetics in early and late middle‐aged adults with uncomplicated type 2 diabetes (T2D). Fifty inactive adults with T2D (16 females) were stratified into late middle‐aged (60–70 years) and early middle‐aged (37–59 years) cohorts and assigned to exercise training or control groups: late middle‐aged training (TR‐LMA, n = 12), late middle‐aged control (CON‐LMA, n = 9), early middle‐aged training (TR‐EMA, n = 17), or early middle‐aged control (CON‐EMA, n = 12). Training groups completed a supervised 12‐week aerobic and resistance programme. V̇O2peak was assessed during incremental cycling while the time constant of the primary phase of oxygen uptake kinetics (τV̇O2p), blood pressure, and cardiac output (inert gas rebreathing) were measured during submaximal cycling before and after the intervention. V̇O2peak increased and τV̇O2p decreased (p < 0.001) in both training groups: TR‐LMA (25.6 ± 5.7 to 28.2 ± 5.3 mL·kg−1·min−1; 39.8 ± 6.5 to 35.0 ± 7.3 s) and TR‐EMA (25.3 ± 5.3 to 28.1 ± 5.3 mL·kg−1·min−1; 43.3 ± 9.0 to 33.8 ± 8.2 s), with no between‐group differences. No changes occurred in CON‐LMA or CON‐EMA, and cardiovascular responses remained unchanged following training. Combined aerobic and resistance training similarly improved V̇O2peak and τV̇O2p in early and late middle‐aged adults with T2D. The improvements in V̇O2 kinetics appear to occur independently of systemic cardiovascular adaptations.Keywords: cardiac output, exercise intolerance, high‐intensity interval training, moderate‐intensity continuous training INTRODUCTION The prevalence of type 2 diabetes (T2D) increases markedly with advancing age, contributing substantially to the growing global burden of the disease. As global populations continue to age, alongside persistent sedentary behavior and rising obesity rates, the clinical and economic burden of T2D is projected to escalate further. Preserving functional capacity and reducing cardiovascular risk across middle age and later life are therefore key clinical priorities in individuals with T2D. In this regard, exercise remains a cornerstone of T2D management because of its well‐established benefits for glycaemic control and cardiovascular risk reduction (Wei et al., 2000). However, individuals with T2D commonly exhibit impaired exercise capacity, which may limit their ability to meet physical activity recommendations (Johnson et al., 2005; Morrato et al., 2007). Specifically, lower peak oxygen uptake (V̇O2peak) is consistently observed across age groups in T2D compared with individuals without the condition (Alvares et al., 2024; Green et al., 2015; Mac Ananey et al., 2011; O'Connor et al., 2012, 2015; Regensteiner et al., 1998). Given that V̇O2peak is a strong predictor of cardiovascular and all‐cause mortality (Wei et al., 2000), reductions in aerobic capacity carry important prognostic implications. Beyond maximal capacity, T2D also alters submaximal exercise responses. Adults with T2D under 60 years demonstrate slower oxygen uptake (V̇O2) kinetics during moderate‐ and high‐intensity exercise, reflected by a prolonged time constant of the primary phase of the V̇O2 response (τV̇O2p) (Bauer et al., 2007; Gildea et al., 2021; Kiely et al., 2015; Mac Ananey et al., 2011; O'Connor et al., 2012; O'Connor et al., 2015), a key indicator of exercise tolerance (Hughson et al., 2001; Poole & Jones, 2012). Slower kinetics increase the oxygen deficit at exercise onset, promote greater reliance on substrate‐level phosphorylation and may contribute to earlier fatigue (Jones & Poole, 2005). Mechanisms influencing these impairments include reduced left ventricular filling (Wilson, Wilkins, et al., 2017; Wilson, Wilson, et al., 2017), impaired peripheral vasodilation and microvascular function (Bauer et al., 2007; Gildea et al., 2019; Gildea et al., 2021; Kiely et al., 2014; Mac Ananey et al., 2011; Poitras et al., 2015; Rocha et al., 2019; Wilkerson et al., 2011), and restricted muscular oxygen extraction (Kelley et al., 2002). Importantly, τV̇O2p appears relatively preserved in adults with T2D in the seventh decade (60–70 years) compared with age‐matched controls, suggesting that age may modify the physiological phenotype of exercise impairment in T2D (O'Connor et al., 2015; Wilkerson et al., 2011). Whether this reflects age‐related ceiling effects (O'Connor et al., 2015) or compensatory cardiovascular and metabolic adaptations (Wilkerson et al., 2011) remains unclear. Short‐term exercise interventions incorporating aerobic or combined aerobic and resistance training typically increase V̇O2peak by approximately 15%–20% (Brandenburg et al., 1999; Gildea et al., 2021a; Macananey et al., 2012; O'Connor et al., 2025; Winding et al., 2018) and accelerate τV̇O2p by 15%–30% in individuals with T2D (Brandenburg et al., 1999; Gildea et al., 2022; Gildea et al., 2021b; Green et al., 2020; Macananey et al., 2012). Such improvements are clinically meaningful, given the strong association between cardiorespiratory fitness and survival. However, most studies examining training‐induced adaptations in V̇O2 kinetics have involved younger or predominantly early middle‐aged cohorts (Brandenburg et al., 1999; Gildea et al., 2021b; Green et al., 2020; Macananey et al., 2012; O'Connor et al., 2025), with no direct comparisons across age groups. Consequently, it remains unclear whether late middle age/young older adults with T2D demonstrate V̇O2peak and V̇O2 kinetics adaptations to structured exercise training that differ from those observed in early middle‐aged individuals with the disease, or whether age influences the magnitude and/or mechanisms of these responses. Given the rising prevalence of T2D in older populations and the central role of exercise in disease management, clarifying age‐specific training responses has important implications for prescription and risk stratification. Accordingly, the present study investigated the effects of structured exercise training on V̇O2peak and V̇O2 kinetics in early versus late middle‐aged adults with T2D. To further determine if changes in oxygen uptake kinetics were associated with central cardiovascular adaptations and enhanced oxygen delivery, cardiac output was assessed during submaximal exercise. MATERIALS AND METHODS Design and participants This investigation formed one component of a larger project examining how sex‐related differences (Green et al., 2020) and treatment with pioglitazone (O'Connor et al., 2025) may influence the exercise training‐induced effects in V̇O2peak and V̇O2 kinetics and cardiovascular responses during submaximal exercise in individuals with T2D. Fifty inactive adults with T2D were stratified by age into “late middle‐aged” (60–70 years; n = 21, 4 females) and “early middle‐aged” (37–59 years; n = 29, 12 females) cohorts. Two females were premenopausal. Inclusion required a confirmed diagnosis of T2D for under 10 years and HbA1c values below 10%. Exclusion criteria comprised insulin use, active smoking, contraindications to exercise testing or training, BMI above 40 kg/m2, and any major cardiovascular, hepatic, or renal disease (Table 1). Recruitment occurred through outpatient departments at St. Columcille's Hospital and St. Vincent's University Hospital in Dublin. TABLE 1Participants' clinical characteristics and activity levels at baseline and changes in physical characteristics and GXT variables before and after the 12‐week intervention in early and late middle‐aged participants with type 2 diabetes.Pre/Post interventionCON‐LMA (n = 9)CON‐EMA (n = 12)TR‐LMA (n = 12)TR‐EMA (n = 17)Clinical characteristics and activity levelsSex (male, female), n 7, 27, 510, 210, 7Age (years)64 ± 3 a 51 ± 363 ± 3 a 54 ± 3Time since diabetes diagnosis (yr)5.2 ± 2.13.2 ± 1.84.5 ± 2.74.9 ± 3.5Diabetes medicationDiet only, n (%)2 (22)2 (17)3 (25)8 (47)Metformin, n (%)7 (78)10 (83)9 (75)8 (47)Sulfonylurea, n (%)1 (11)1 (8)3 (25)4 (24)Glucagon‐like peptide‐1, n (%)1 (11)1 (8)Thiazolidinediones, n (%)1 (11)10 (83)Antihypertensive medication, n (%)6 (67)5 (42)7 (58)8 (47)Statins, n (%)7 (78)5 (42)8 (67)8 (47)Habitual physical activityLPA, h/day4.4 ± 2.13.9 ± 2.34.3 ± 0.84.6 ± 2.0MVPA, h/day1.0 ± 0.60.8 ± 0.60.9 ± 0.61.1 ± 0.7Changes in physical characteristics and GXT variablesBody Mass (kg)Pre94.4 ± 11.989.5 ± 12.490.0 ± 12.686.9 ± 14.4Post94.0 ± 12.489.8 ± 13.490.3 ± 12.186.7 ± 15.3BMI (kg.m−2)Pre31.6 ± 3.531.7 ± 4.329.4 ± 2.431.6 ± 4.4Post31.5 ± 3.631.9 ± 4.329.6 ± 2.431.5 ± 4.5HbA1c (%)Pre6.4 ± 0.76.3 ± 0.67.0 ± 1.16.7 ± 0.7Post6.4 ± 0.76.4 ± 0.47.0 ± 1.26.6 ± 0.7Glucose (mM)Pre7.3 ± 1.26.8 ± 1.37.9 ± 1.67.3 ± 1.1Post7.8 ± 1.47.5 ± 2.18.2 ± 2.37.0 ± 1.2V̇O2peak (l.min−1) c , d Pre2.19 ± 0.452.16 ± 0.562.31 ± 0.612.19 ± 0.51Post2.11 ± 0.492.22 ± 0.672.57 ± 0.63 b 2.43 ± 0.57 b V̇O2peak (mL.min−1.kg−1) c , d Pre23.4 ± 4.624.1 ± 5.425.6 ± 5.725.3 ± 5.3Post22.7 ± 5.024.7 ± 6.328.2 ± 5.3 b 28.1 ± 5.3 b POpeak (W) c , d Pre142 ± 27134 ± 38153 ± 47136 ± 32Post135 ± 30138 ± 43178 ± 51 b 161 ± 40 b HRpeak (beat.min−1)Pre146 ± 15159 ± 13155 ± 17153 ± 15Post145 ± 13163 ± 17157 ± 20149 ± 14GET (W) c , d Pre114 ± 25111 ± 29123 ± 34115 ± 27Post112 ± 27116 ± 33146 ± 40 b 138 ± 35 b O2 pulse (mL.beats−1) c , d Pre16.1 ± 3.215.3 ± 3.616.6 ± 3.116.8 ± 3.7Post15.7 ± 3.715.3 ± 4.318.1 ± 2.7 b 18.9 ± 3.5 b Note: Data are mean ± SD; n number of participants.Abbreviations: BMI, body mass index; CON, non‐exercising control; EMA, early middle‐aged; GET, gas exchange threshold; HbA1c, glycated hemoglobin; HR, heart rate; LMA, late middle‐aged; LPA, light‐intensity physical activity; MVPA, moderate‐to‐vigorous physical activity; PO, power output; TR, exercise training; V̇O2, oxygen consumption. a Significantly different than both early middle‐aged groups (CON‐EMA and TR‐EMA) (p < 0.05). b Significantly different than pre‐training within the same group (p < 0.05). c Significant main effect of time (p < 0.05). d Significant time × group interaction (p < 0.05). Participants' clinical characteristics and activity levels at baseline and changes in physical characteristics and GXT variables before and after the 12‐week intervention in early and late middle‐aged participants with type 2 diabetes. Note: Data are mean ± SD; n number of participants. Abbreviations: BMI, body mass index; CON, non‐exercising control; EMA, early middle‐aged; GET, gas exchange threshold; HbA1c, glycated hemoglobin; HR, heart rate; LMA, late middle‐aged; LPA, light‐intensity physical activity; MVPA, moderate‐to‐vigorous physical activity; PO, power output; TR, exercise training; V̇O2, oxygen consumption. Significantly different than both early middle‐aged groups (CON‐EMA and TR‐EMA) (p < 0.05). Significantly different than pre‐training within the same group (p < 0.05). Significant main effect of time (p < 0.05). Significant time × group interaction (p < 0.05). Participants first underwent a treadmill electrocardiogram stress assessment using the Bruce protocol and were non‐randomly assigned into one of four study groups: (1) late middle‐aged control (CON‐LMA: n = 9, 2 females, 64 ± 3 years), (2) early middle‐aged control (CON‐EMA: n = 12, 5 females, 51 ± 3 years); (3) late middle‐aged exercise training (TR‐LMA: n = 12, 2 females, 63 ± 3 years); and (4) early middle‐aged exercise training (TR‐EMA: n = 17, 7 females, 54 ± 3 years). Those in the exercise arms completed a 12‐week supervised training program, whereas control participants did not take part in any structured exercise during the study period. The non‐random allocation primarily reflected differences in participants' ability to travel to the training facility. All volunteers provided written informed consent. Ethical approval was granted by Trinity College Dublin (Ref: 131003) and St. Vincent's Healthcare. The study adhered to the guidelines set forth in the Declaration of Helsinki. Exercise training intervention The exercise intervention comprised up to 36 sessions over 12 weeks (three sessions per week). Adherence was defined as attendance at ≥32 sessions. All sessions were supervised by members of the research team and involved combined aerobic and resistance (concurrent) training, as this approach has been shown to produce greater reductions in HbA1c in individuals with T2D than either modality alone (Church et al., 2010). Further details of the training phases are provided elsewhere (O'Connor et al., 2025). Briefly, each session lasted 60–90 min and consisted of five components: (1) continuous cycling for 20–25 min at 70%–80% HRmax; (2) upper and lower body flexibility exercises; (3) core strengthening exercises (sit‐ups, planks, and back extensions with 20–30 repetitions each); (4) resistance training for upper (bench press, upright row, shoulder press, lat pulldown) and lower body (leg extension, leg curl, leg press) at 60%–70% one‐repetition maximum for two sets of 10 reps; and (5) high‐intensity interval exercise consisting of up to three 90–120 s bouts at ~90% HRmax separated by two‐minute recovery periods. Training intensity and/or duration was progressively increased after weeks 4 and 8. Testing Participants attended the laboratory on two occasions both before and after the 12‐week intervention. All testing sessions were conducted at the same time of day and separated by 72 h. On the first visit, participants performed a graded exercise test (GXT) to volitional exhaustion on a cycle ergometer. The test started at 40 W and the workload increased by 30 W every 3 min until the required cadence could no longer be maintained (Egana et al., 2007). The gas exchange threshold (GET) was identified using the V‐slope method (Beaver et al., 1986), peak responses (including V̇O2) were defined as the highest 30‐s averaged values, and peak power output was calculated as previously described (Balmer et al., 2000). During the second visit, participants completed six constant‐load cycling bouts. Each bout consisted of 3 min of ‘unloaded’ cycling (10 W) followed by 6 min at a workload corresponding to 80% of the individually determined GET. After the intervention, workloads were adjusted to reflect any change in power output at GET, ensuring the same relative intensity (80% GET) before and after training. Bouts were separated by 12‐min recovery periods to allow HR and blood lactate (measured in a subgroup, n = 25) to return to baseline. Oxygen uptake and HR were recorded continuously during the first four bouts, whereas CO and mean arterial pressure (MAP) were measured during the final two bouts. Previous work indicates that V̇O2 kinetic estimates are comparable whether repeated bouts are performed within a single day or across multiple days (Spencer et al., 2011). Venous blood samples were obtained after a 12‐h overnight fast 24–48 h prior to exercise testing. Participants were instructed to avoid caffeine and alcohol for 24 h before each laboratory visit. Measurements Pulmonary gas exchange was assessed breath‐by‐breath using a system (Innocor, Innovision A/S, Odense, Denmark) which measured expired air flow and respiratory gases sampled at the mouthport. This system computed V̇O2, rate of CO2 output (V̇CO2), minute ventilation (V̇E), and respiratory exchange ratio (RER). CO was measured using the same system but with the rebreathing of an inert gas mixture (sulfur hexafluoride and nitrous oxide). Cardiac output measurements were restricted to rest, 30 and 240 s of exercise because of the time required for adequate rebreathing and measurement (Jakovljevic et al., 2008; Mac Ananey et al., 2011). Heart rate was recorded every 5 s (S610i, Polar Electro Oy, Finland). Stroke volume (SV) was calculated from simultaneous HR and CO measurements. MAP was estimated from sphygmomanometric measurements of brachial arterial pressures (0.33 systolic + 0.66 diastolic) performed at the same times CO was measured. Systemic vascular conductance (SVC) was calculated from CO and MAP measurements (SVC = CO/MAP). All cardiovascular measurements were mean values of responses during two exercise bouts. Analysis of kinetic responses of V̇O2 and cardiovascular variables Analysis of V̇O2 kinetics was performed using data from the first four submaximal cycling bouts completed at a relative intensity of 80% GET before and after the intervention. Baseline V̇O2 was defined as the mean value during the final 60 s of unloaded cycling preceding the transition in workload. Breath‐by‐breath data from 0 to 240 s following the transition were interpolated to 1‐s intervals, temporally aligned, averaged across bouts, and smoothed with a five‐second moving average (Keir et al., 2014). The processed responses were fitted using either a monoexponential (Equation 1) or biexponential (Equation 2) model according to the response profile. Most responses (83%) were adequately described by a monoexponential function representing the primary adjustment phase. When a distinct slow component was present (17% of cases; mean ± SD amplitude: 110 ± 73 mL·min−1), a biexponential model was applied. The equations are: In the model equations, a represents baseline V̇O2 at 10 W; A₁ and A2 denote the amplitudes of the primary and slow phases; τ₁ and τ 2 their respective time constants; and TD₁ and TD2 the associated time delays. Conditional terms F₁ and F2 ensure each phase is fitted only after its delay. To minimize the influence of the cardiodynamic phase, the first 20 s of data were excluded, although TD₁ was allowed to vary during fitting to improve parameter estimation. Because T2D primarily affects the primary phase, only τ₁ (τV̇O2p) and derived variables are reported. Model fitting used a two‐step weighted least‐squares non‐linear regression procedure with removal of outliers after the initial step. V̇O2 gain (A₁/Δpower) and end‐exercise V̇O2 (mean of the final 30 s) were also calculated. Dynamic responses of CO, SV, MAP, SVC, and HR were derived from the averaged responses during the final two exercise bouts. Dynamic responses were quantified by expressing the change observed during the first 30 s of exercise as a percentage of the total change measured over 240 s (Δ30/Δ240 × 100%). Statistical analyses Age, duration of diabetes, physical activity levels, baseline physical characteristics, and GXT‐derived variables were initially compared between groups using one‐way ANOVA. Changes in physical characteristics, peak physiological responses, V̇O2 kinetic parameters, and cardiovascular dynamic variables were analyzed using a two‐factor mixed ANOVA with factors for time (pre‐training vs. post‐training) and group (CON‐LMA, CON‐EMA, TR‐LMA, TR‐EMA). When significant effects were detected, Bonferroni‐adjusted post hoc tests were used to identify differences between groups. Differences in the magnitude of training‐induced changes between the exercise groups were further examined using independent t‐tests. Statistical significance was accepted at p < 0.05. All data are presented as mean ± SD. RESULTS Participant characteristics and GXT variables Participant characteristics and responses from the GXT are presented in Table 1, and individual and mean group changes in V̇O2peak are illustrated in Figure 1. Adherence to the exercise programme was high in both training groups (TR‐LMA, TR‐EMA), with participants completing an average of 35 sessions (range: 32–36) out of a possible 36. Participants in both early middle‐aged groups were younger than in the late middle‐aged groups (p < 0.001). Duration of diabetes and habitual physical activity levels were similar across groups (all p > 0.29). Baseline physical characteristics and GXT‐derived variables were also not different among groups (all p > 0.06). Body mass and HbA1c remained unchanged in all groups following the 12‐week intervention. In contrast, significant increases were observed in absolute and relative V̇O2peak, peak power output (POpeak), O2 pulse, and power output at GET in both exercise groups after training (for all variables, main effect of time, p < 0.001 and significant time × group interaction, p < 0.001). The magnitude of these improvements, expressed both in absolute terms and as percentages, did not differ between TR‐LMA and TR‐EMA (t‐tests, all p > 0.2). Individual and mean responses of peak oxygen uptake (V̇O2peak, panels A–D) and time constant of the primary V̇O2 response (τV̇O2p, panels E–H) in the four experimental groups (Late middle‐aged control, n = 9; Early middle‐aged control, n = 12; Late middle‐aged training, n = 12; Early middle‐aged training, n = 17) before and after the 12‐week intervention. Error bars represent SD. A two‐factor mixed ANOVA was used to carry out the analyses. * Significantly different than pre‐training within the same group (p < 0.05). VO2 kinetics Estimates of the V̇O2 kinetic parameters and derived measurements for the primary phase are summarized in Table 2 together with the statistical results, while individuals and mean changes of the time constant of the primary V̇O2 response (τV̇O2p) are shown in Figure 1. Following the intervention, τV̇O2p was reduced, indicating faster kinetics in both exercise groups only (significant time × group interaction, p < 0.001). The % magnitude of this improvement, although numerically larger in TR‐EMA compared with TR‐LMA (~21% vs. ~12% respectively), was not significantly different between the groups (t‐tests, p = 0.11). Significant main effects of time were also observed for the primary amplitude of V̇O2 and end‐exercise V̇O2 (p = 0.005 and p = 0.003, respectively). Although these increases were mainly observed in the exercise groups, the time × group interactions were not statistically significant (p = 0.20 and p = 0.14). TABLE 2V̇O2 kinetics parameters and changes in cardiovascular variables during submaximal exercise before and after a 12‐week intervention in early and late middle‐aged male and female participants with type 2 diabetes.Pre/Post interventionCON‐LMA (n = 9)CON‐EMA (n = 12)TR‐LMA (n = 12)TR‐EMA (n = 17) V̇O 2 kinetics parameters V̇O2 base (l.min−1)Pre0.70 ± 0.100.72 ± 0.190.81 ± 0.130.68 ± 0.19Post0.72 ± 0.110.72 ± 0.170.83 ± 0.180.73 ± 0.17τV̇O2p (s) b , c Pre43.8 ± 10.139.3 ± 11.239.8 ± 6.543.3 ± 9.0Post44.8 ± 9.040.2 ± 10.935.0 ± 7.3 a 33.8 ± 8.2 a V̇O2 TD (s)Pre16.6 ± 8.614.8 ± 7.820.3 ± 9.315.6 ± 8.0Post15.9 ± 9.515.1 ± 6.519.3 ± 6.818.9 ± 5.3V̇O2 Amp (l.min−1) b Pre0.82 ± 0.280.82 ± 0.260.96 ± 0.340.88 ± 0.27Post0.85 ± 0.160.83 ± 0.311.08 ± 0.351.02 ± 0.29V̇O2 Gain (mL.min−1.W−1)Pre10.0 ± 1.710.4 ± 1.211.3 ± 3.410.8 ± 2.0Post11.1 ± 1.410.0 ± 1.410.3 ± 1.210.2 ± 1.3End V̇O2 (l.min−1) b Pre1.56 ± 0.301.56 ± 0.401.79 ± 0.401.56 ± 0.39Post1.59 ± 0.241.61 ± 0.481.95 ± 0.481.79 ± 0.35 Dynamic changes in cardiovascular variables ΔCO30/ΔCO240 (%)Pre67 ± 1571 ± 1572 ± 1264 ± 14Post72 ± 1269 ± 1873 ± 869 ± 11ΔHR30/ΔHR240 (%)Pre53 ± 1048 ± 1550 ± 1850 ± 18Post48 ± 1451 ± 2550 ± 852 ± 11ΔSV30/ΔSV240 (%)Pre99 ± 46105 ± 3197 ± 3290 ± 27Post117 ± 3699 ± 3489 ± 33103 ± 32ΔMAP30/ΔMAP240 (%)Pre62 ± 2965 ± 1268 ± 1572 ± 23Post73 ± 1878 ± 1370 ± 1967 ± 13ΔSVC30/ΔSVC240 (%)Pre70 ± 3079 ± 2177 ± 1266 ± 24Post76 ± 1671 ± 2282 ± 2372 ± 14 Note: Data are mean ± SD. Cardiovascular responses over 30 s are expressed as percentages of responses over 240 s of exercise (i.e., change from rest) to provide estimates of the dynamic responses of these variables during exercise.Abbreviations: Amp, amplitude; base, baseline; CO, cardiac output; CON, non‐exercising control; EMA, early middle‐aged; End; end‐exercise; GET, gas exchange threshold; HR, heart rate; LMA, late middle‐aged; MAP, mean arterial pressure; SV, stroke volume; SVC, systemic vascular conductance; TD, time delay; TR, exercise training; τpV̇O2, time constant of V̇O2 during the primary phase. a Significantly different than pre‐training within the same group (p < 0.05). b Significant main effect of time (p < 0.05). c Significant time × group interaction (p < 0.05). V̇O2 kinetics parameters and changes in cardiovascular variables during submaximal exercise before and after a 12‐week intervention in early and late middle‐aged male and female participants with type 2 diabetes. Note: Data are mean ± SD. Cardiovascular responses over 30 s are expressed as percentages of responses over 240 s of exercise (i.e., change from rest) to provide estimates of the dynamic responses of these variables during exercise. Abbreviations: Amp, amplitude; base, baseline; CO, cardiac output; CON, non‐exercising control; EMA, early middle‐aged; End; end‐exercise; GET, gas exchange threshold; HR, heart rate; LMA, late middle‐aged; MAP, mean arterial pressure; SV, stroke volume; SVC, systemic vascular conductance; TD, time delay; TR, exercise training; τpV̇O2, time constant of V̇O2 during the primary phase. Significantly different than pre‐training within the same group (p < 0.05). Significant main effect of time (p < 0.05). Significant time × group interaction (p < 0.05). Cardiovascular dynamics Group mean cardiovascular responses are illustrated in Figure 2. Neither the exercise intervention nor the control period significantly altered the rate of adjustment of any cardiovascular variable (Table 2). However, significant main effects of time were observed for CO, SV, and SVC at 30 s and also for SV and SVC at 240 s (all p < 0.05), largely reflecting the higher exercise workloads achieved after training in the exercise groups. No significant time × group interactions were detected (all p > 0.06). At baseline, there was a main effect of time for SV (p < 0.05) while mean arterial pressure (MAP) was lower (p = 0.03) post‐intervention in the CON‐EMA group (group × time interaction, p = 0.045). Mean ± SD cardiovascular responses at rest, 30 and 240 s of exercise before (open circles) and after (closed circles) the 12‐week intervention in the four experimental groups (Late middle‐aged control, n = 9; Early middle‐aged control, n = 12; Late middle‐aged training, n = 12; Early middle‐aged training, n = 17). Statistical outcomes are further described within the results section. CO, cardiac output; HR, heart rate; SV, stroke volume; MAP, mean arterial pressure; SVC, systemic vascular conductance. * Significantly different than pre‐training within the same group (p < 0.05). DISCUSSION To our knowledge, this is the first study to examine exercise training‐induced changes in cardiorespiratory fitness and oxygen uptake kinetics across two middle‐aged cohorts of individuals with T2D and mild obesity. The main findings were: (1) both exercise groups (TR‐LMA and TR‐EMA) showed significant improvements in V̇O2peak and τV̇O2p, with no differences in the magnitude of adaptation between groups; and (2) the improvement in V̇O2 kinetics was not accompanied by alterations in cardiovascular dynamics. Taken together, these findings suggest that exercise training enhances both maximal and submaximal exercise tolerance to a magnitude not different between previously untrained “early middle‐aged” (40s–50s) and “late middle‐aged” (60s) individuals with T2D of similar disease duration and baseline physiological profile, and that the observed improvements in submaximal responses appear to occur independently of systemic cardiovascular adaptations. Peak cardiorespiratory fitness Both exercise groups exhibited clinically meaningful increases in V̇O2peak (~12%). Given that V̇O2peak is a strong independent predictor of cardiovascular and all‐cause mortality (Ross et al., 2016), such improvements are likely to translate into meaningful reductions in risk, particularly in late middle‐aged adults with T2D who exhibit elevated baseline cardiometabolic burden. The concomitant rise in O2 pulse, improved V̇O2peak without a change in HRmax, supports the presence of true physiological adaptation rather than differences in effort or test familiarization. Mechanistically, exercise‐induced improvements in V̇O2peak in individuals with T2D, relative to non‐exercising controls, have been attributed to both central and peripheral adaptations. Central adaptations may include increased left ventricular end‐diastolic volume and stroke volume, mediated by improved ventricular compliance and preload (Wilson et al., 2019). Peripheral mechanisms likely include enhanced microvascular oxygen delivery (Gildea et al., 2021b), improved vascular function (Schreuder et al., 2015), structural vascular remodeling (Madsen et al., 2015), and an expanded capillary‐to‐myocyte interface (Mortensen et al., 2019), potentially mediated by an increase in the number of red blood cell–perfused capillaries within skeletal muscle (Padilla et al., 2006). Our laboratory previously showed that peak V̇O2 was reduced in participants with T2D aged <60 years (by ~22%) and in those aged 60–70 years (by ~12%) compared with their age‐matched healthy counterparts (O'Connor et al., 2015). The current study, which included participants with comparable disease duration and clinical characteristics, extends these observations by demonstrating that impairments in V̇O2peak remain modifiable in these age groups with uncomplicated T2D. Oxygen uptake kinetics during moderate‐intensity exercise In addition to improvements in maximal capacity, both groups exhibited significant acceleration of V̇O2 kinetics during moderate‐intensity exercise (~12–21% reduction in τV̇O2p). Given that τV̇O2p reflects the efficiency of the transition from rest to activity, these findings have clinical relevance. Faster kinetics reduce the oxygen deficit at exercise onset, limit reliance on substrate‐level phosphorylation, and may delay fatigue, thereby improving tolerance to everyday physical tasks, an effect particularly relevant for adults in late middle age who may be at increased risk of functional decline. Although not directly established in T2D, evidence from other clinical populations, such as heart failure and chronic obstructive pulmonary disease, suggests that training‐induced changes in τV̇O2p are likely mediated by alterations in the rate of oxygen delivery to, and/or utilization by, the contracting skeletal muscle (Hughson, 2009; Poole & Jones, 2012). To examine potential central contributions, we assessed early (30 s) versus later (240 s) changes in cardiovascular variables during submaximal exercise. Consistent with recent findings (O‘Connor et al., 2025; Green et al., 2020), no significant alterations in systemic cardiovascular dynamics were detected following training, suggesting that large changes in central haemodynamic kinetics were unlikely to explain the observed acceleration in τV̇O2p. Our previous baseline study suggested that slowed τV̇O2p in middle‐aged males with T2D was at least partly attributable to blunted early SVC (estimated from the slope MAP and CO) responses and impaired muscle perfusion (O'Connor et al., 2015). The exaggerated pressor response during submaximal exercise in middle‐aged T2D participants, despite smaller rises in CO, implied reduced vascular conductance in contracting skeletal muscle. In contrast, older males with T2D did not exhibit this additional vascular impairment, consistent with the absence of further slowing in τV̇O2p beyond that attributable to aging alone (O'Connor et al., 2015). Others also reported unaltered τV̇O2p in older males with T2D compared with controls alongside a greater increase in muscle deoxygenation per unit of V̇O2, suggesting enhanced O2 extraction during exercise as a compensatory response to impaired muscle blood flow (Wilkerson et al., 2011). The present findings extend these observations by demonstrating that, despite differences in baseline exercise responses reported across age groups, structured exercise training improves dynamic oxygen uptake in both early and late middle‐aged adults with T2D without altering early SCV responses. This suggests that the training‐induced acceleration of V̇O2 kinetics is more likely driven by enhanced peripheral oxygen delivery–utilization matching rather than improvements in central oxygen delivery although peripheral adaptations were not measured in the current investigation. Such peripheral adaptations may include improved local matching of oxygen delivery to muscle metabolic demand (Gildea et al., 2021b) enhanced vasodilatory capacity, and improved dynamic microvascular responses (Green et al., 2022). Clinical implications Despite the increasing prevalence of T2D across middle and later life and elevated cardiovascular risk associated with the condition, responsiveness to exercise training, reflected in improvements in both maximal and submaximal exercise capacity, appears to be maintained across early and late middle age, with no clear evidence of reduced physiological plasticity. Overall, the data support the use of structured aerobic or combined training programmes as an effective approach to improve key functional physiological markers in this population. Limitations This study has several limitations that should be acknowledged. First, due to the limited temporal resolution of the inert gas rebreathing technique, cardiac output (CO) was not assessed continuously; therefore, subtle alterations in dynamic central haemodynamics cannot be excluded. Future studies employing higher‐resolution methodologies may help to further delineate the relative contributions of central and peripheral adaptations. Second, the proportion of males and females differed between groups. However, recent evidence indicates that adaptations in V̇O2peak and V̇O2 kinetics following exercise training are not influenced by sex in individuals with type 2 diabetes (Green et al., 2020), supporting the inclusion of both sexes in the present study. Furthermore, excluding female participants from the analysis did not alter the main outcomes (data not shown). Nevertheless, the unequal sex distribution may have influenced baseline physiological characteristics, so future studies specifically investigating the age‐related differences in exercise adaptations should aim to recruit more balanced sex distributions across groups. Third, the proportion of participants managed with diet alone was numerically higher in the early middle‐aged training group, which may have influenced the study outcomes. Fourth, participants had relatively well‐controlled diabetes (mean HbA1c approximately 6.3%–7.0%), which may limit the generalisability of our findings to individuals with poorer glycaemic control or more advanced disease, who typically exhibit greater impairments in cardiorespiratory fitness and exercise capacity (Alvares et al., 2024; Soares & Lessard, 2024). Fifth, a concurrent training programme was chosen because of its high ecological validity in type 2 diabetes. However, the present study cannot determine the relative contributions of the aerobic and resistance components to the observed improvements in V̇O2peak and τV̇O2p. Although the aerobic component was based on a previously published protocol that elicited improvements in systemic cardiovascular dynamics in adults with type 2 diabetes, no such changes were observed in the present study. Therefore, it cannot be excluded that the aerobic training stimulus was insufficient to elicit detectable central cardiovascular adaptations. Finally, the mean age difference between the early and late middle‐aged groups was relatively modest (~10 years). This limited age separation restricts the extent to which the present findings can be interpreted as evidence of age‐related differences in exercise trainability and limits their generalisability across a broader age range. Nevertheless, the present study provides a relevant benchmark for the expected magnitude of exercise‐training adaptations in individuals with T2D in their 40s–50s and 60s, despite previous evidence of impaired V̇O2 kinetics in those in their 40s–50s, but not in those in their 60s, relative to their respective age‐matched non‐diabetic controls (O'Connor et al., 2015; Wilkerson et al., 2011). CONCLUSION In conclusion, 12 weeks of structured exercise training significantly improved V̇O2peak and accelerated V̇O2 kinetics in early and late middle‐aged adults with uncomplicated T2D and same disease duration. These adaptations were similar between age groups and were not accompanied by changes in systemic cardiovascular dynamics. The findings reinforce the clinical value of exercise training across early and late middle age in adults with T2D and highlight preserved physiological adaptability within these age groups. Further investigation is needed to determine whether comparable benefits extend to individuals at more advanced ages, including those over 70 years and/or with longer disease duration. AUTHOR CONTRIBUTIONS Catherine Kiely: Conceptualization; formal analysis; investigation; methodology. Eamonn O'Connor: Conceptualization; formal analysis; investigation; methodology. Donal O'Shea: Conceptualization; funding acquisition; investigation; methodology; resources. Simon Green: Conceptualization; funding acquisition; investigation. Mikel Egaña: Conceptualization; formal analysis; funding acquisition; methodology; project administration; supervision. FUNDING INFORMATION This research was supported by Research Ireland (Grant No. 08/RFP/BMT1342). CONFLICT OF INTEREST STATEMENT The authors declare no conflicts of interest. DATA AVAILABILITY STATEMENT Data will be made available on request within the limitations of the informed consent. Requests should be made to the corresponding author. REFERENCES Associated Data Data Availability Statement Data will be made available on request within the limitations of the informed consent. Requests should be made to the corresponding author.