The impact of online-delivered controlled physical activity on cardiorespiratory fitness and heart rate variability in breast cancer survivors

Abstract

Breast cancer survivors are at increased risk of cardiovascular and autonomic dysfunction following adjuvant chemotherapy, and exercise interventions may help mitigate these effects. This randomized prospective study evaluated the effects of a supervised, online-delivered exercise intervention on cardiorespiratory fitness, cardiac autonomic modulation, body composition, and quality of life in breast cancer survivors after chemotherapy. Seventy-two women were allocated to an exercise group or usual care. The intervention consisted of a 12-week home-based exercise program conducted online, with aerobic intensity individualized to 60–80% of peak oxygen uptake. Cardiorespiratory fitness was assessed using cardiopulmonary exercise testing and the 6-min walk test, while heart rate variability was monitored longitudinally and analyzed using linear mixed-effects models. Body composition and patient-reported outcomes were assessed at baseline, post-chemotherapy, and post-intervention. The online supervised exercise resulted in a significant improvement in peak oxygen uptake (+ 2.1 ml kg⁻¹ min⁻¹), which exceeded changes observed in the control group (p = 0.009). Improvements in 6-min walk distance and recovery of heart rate variability occurred irrespective of group allocation. Changes in body composition, quality of life and the improvement of heart rate variability metrics did not differ between groups.

Introduction

Breast cancer survivors often face complex adverse effects following primary oncological treatment that not only compromise their daily functioning but also increase the risk of cardiovascular and metabolic comorbidities. These risks are further enhanced by prolonged exposure to adjuvant endocrine treatment, which contributes to weight gain and unfavorable body composition.

Cardiovascular compromise in breast cancer survivors is mediated through multiple pathways. While anthracycline-based cytotoxic regimens may induce dose-dependent myocardial injury leading to measurable reductions in the left ventricular ejection fraction, taxane therapy has been linked to adverse cardiac outcomes through its detrimental impact on autonomic nervous system^1–3. Some studies suggest that the imbalance in heart rate regulatory system may precede overt clinical cardiotoxicity and serve as sensitive predictor of long-term cardiovascular risk and all-cause mortality in oncology populations^4–6. The compromise of neurocardiac function manifests through the loss of normal heart rate variability (HRV) which reflects the interplay between sympathetic and parasympathetic nervous system inputs to the sinoatrial node. Technically, it measures oscillations in the heart-beat interval time series that occur at different frequencies. High-frequency (HF) oscillations primarily index vagal modulation related to respiratory sinus arrhythmia, while low-frequency (LF) components reflect a combination of sympathetic and baroreflex influences^7,8. Total HRV variance indicates the overall responsiveness of the autonomic regulation and provides complementary information on neurocardiac flexibility. Mean of the sum of squared differences (MSSD) is another HRV metric that captures rapid changes in RR intervals occurring from one heartbeat to the next and is of particular significance in orthostatic testing, or repeated measurements across days⁷. Declines in these parameters signify impaired neurocardiac regulation and diminished physiological adaptability.

Exercise interventions represent a widely accepted strategy to improve both cardiopulmonary function and vagal reactivation—each a vital component of cardiovascular resilience. Beyond improvements in functional capacity, controlled physical activity exerts its benefits in cancer survivors via extensive molecular signaling. Although meta analyses suggest that physical activity does not significantly alter specific tumor-targeting immune cell activity^9,10, multiple effects of exercise on immune cells have been reported in the literature. Moreover, contracting skeletal muscle releases a broad spectrum of biologically active molecules termed myokines, which together with other exerkines produced by adipose tissue, liver, immune system, and the cardiovascular tissue contribute to immune modulation, metabolic reprogramming, and mitochondrial adaptation—processes highly relevant in breast cancer survivors exposed to chemotherapy and endocrine treatment^11–13. Within this framework, improvements in cardiorespiratory fitness and autonomic regulation may be understood as downstream integrative manifestations of these systemic adaptations rather than isolated physiological effects of training.

Despite the evidence supporting the benefits of physical activity in survivorship care^14–17, exercise prescriptions across clinical studies remain inconsistent. Current general guidelines recommend at least 150 min per week of moderate-intensity activity or 75–150 min per week of vigorous-intensity, combined with resistance exercise at least twice a week^18–20.

A substantial body of research has explored various exercise intensity prescriptions, mostly tailoring the training load to an individual’s cardiorespiratory fitness^21,22. Intensity modulation based on real-time heart rate monitoring or ventilatory thresholds enables precise control of exertion, aligning the intervention with each patient’s cardiovascular adaptations without excessive physiological strain. Notably, aerobic training within 60–80% of an individual’s peak oxygen uptake (VO₂peak) has consistently been associated with significant improvements in cardiorespiratory fitness^23–26. This training range corresponds to moderate-to-vigorous physical activity levels commonly recommended in current exercise oncology guidelines^18,19. However, dose–response relationships between exercise intensity and physiological outcomes in cancer survivors remain incompletely understood, with substantial inter-individual variability in training responses and clinical effects reported across clinical studies.

To address this gap and advance integrated survivorship care, we conducted a prospective study evaluating the effects of a supervised, controlled online exercise intervention program on a spectrum of clinically relevant outcomes. These encompassed autonomic nervous system function (HRV parameters), cardiorespiratory performance (VO₂peak and 6-min walk distance), quality of life (EORTC QLQ-C30), and body composition (BMI, body fat percentage, lean body mass).

Results

Between October 2021 and June 2024, we enrolled 167 women, of these, 72 completed the study with all required measurements and data, 35 in the supported adapted physical activity (SAPA) and 37 in the control (CO) group. The detailed flowchart of patients along with counts can be found in Fig. 1. Of 85 patients allocated to the SAPA group, 50 patients (58.8%) were excluded from the final analysis, with the most common reason for withdrawal being private matters and scheduling conflicts (n = 20), incomplete measurements or lack of response to follow-up (n = 18), a change in treatment strategy rendering the patient ineligible (n = 6) and deterioration of health status (n = 4). The attrition rate in the control group was 45.1% with the most common reasons for early termination being incomplete measurements and lack of response (n = 20) and personals reasons or scheduling conflicts (n = 16). The baseline demography and tumor characteristics were similar between the groups as shown in Table 1.

Fig. 1.

The flow chart of patient enrollment and randomization.

Table 1.

Basic descriptive statistics (N = 72).

Variable	Category	Experimental group (n = 35)	Control group (n = 37)	p value
Age		51.77 ± 10.01 (44.00; 58.00)	56.57 ± 12.52 (47.00; 67.00)	0.101
BMI		25.64 ± 5.24 (21.79; 28.24)	26.17 ± 4.58 (23.18; 28.02)	0.620
Stage	I	14 (40.0%)	14 (37.8%)	0.907
	II	16 (45.7%)	16 (43.2%)
	III	5 (14.3%)	7 (18.9%)
	IV	0 (0.0%)	0 (0.0%)
ECOG	0	32 (91.4%)	31 (83.8%)	0.480
	1	3 (8.6%)	6 (16.2%)
Treatment duration (months)		4.87 ± 1.30 (3.25; 6.00)	4.65 ± 1.43 (3.00; 6.00)	0.567
Radiotherapy	No	5 (14.3%)	5 (13.5%)	1.000
	Yes	30 (85.7%)	32 (86.5%)
Herceptin treatment	No	27 (77.1%)	31 (83.8%)	0.559
	Yes	8 (22.9%)	6 (16.2%)
Hormone treatment	No	15 (42.9%)	14 (37.8%)	0.810
	Yes	20 (57.1%)	23 (62.2%)
VO2peak after chemo (ml kg−1 min−1)		22.05 ± 5.47 (18.03; 25.00)	20.68 ± 4.32 (17.58; 23.51)	0.313

Continuous variables are expressed as mean ± standard deviation and as median with interquartile range. Differences between the experimental and control groups are represented by p values obtained using the Mann–Whitney U test.

Categorical variables are expressed as absolute and relative numbers, and differences between groups were tested using Fisher’s exact test (or Fisher’s exact test with Monte Carlo simulation and 10 000 permutations for the Stage variable).

Effect of controlled physical activity on cardiorespiratory fitness

Participants completed on average 28.1 ± 5.3 out of 36 prescribed sessions, corresponding to a mean attendance of 78.01%. Among completed sessions, 83.44% were performed within the prescribed heart rate intensity range, whereas 16.6% fell below the target zone. The mean perceived exertion across participants was 12.79 ± 1.1 on the Borg RPE scale (range 11.4–14.3), indicating overall adherence to moderate-to-vigorous exercise intensity. A time plot of session attendance across the intervention period is provided in Fig. 2.

Fig. 2.

A time plot of session attendance.

The dynamics of cardiopulmonary parameters investigated throughout the study are shown in Table 2. There was not a significant deterioration of cardiorespiratory fitness observed in any of the groups after the chemotherapy.

Table 2.

Comparison of selected parameters before and after chemotherapy and intervention/recovery, stratified by group.

Variable	Experimental group (n = 35)					Control group (n = 37)
	Baseline (T0)	After chemotherapy (T1)	After final assessments (T2)	p value (T0 vs T1)	p value (T1 vs T2)	Baseline (T0)	After chemotherapy (T1)	After final assessments (T2)	p value (T0 vs T1)	p value (T1 vs T2)
6-min walk distance	559.34 ± 73.37 570.00 (526.75; 606.50)	567.76 ± 68.36 587.00 (512.50; 604.75)	600.01 ± 70.04 605.00 (569.75; 648.75)	0.399	<0.001	533.68 ± 85.91 553.00 (475.00; 600.00)	527.91 ± 86.54 543.00 (480.00; 585.00)	554.12 ± 78.75 570.00 (499.00; 600.00)	0.670	0.001
eVO2 peak	17.81 ± 1.69 18.06 (17.06; 18.90)	18.01 ± 1.57 18.45 (16.74; 18.86)	18.75 ± 1.61 18.86 (18.05; 19.87)	0.399	<0.001	17.20 ± 1.97 17.44 (15.87; 18.75)	17.06 ± 1.98 17.44 (15.99; 18.31)	17.69 ± 1.81 18.06 (16.43; 18.77)	0.670	< 0.001
Resting heart rate	82.06 ± 14.47 78.00 (71.00; 93.50)	83.00 ± 13.17 82.00 (73.50; 92.00)	82.91 ± 10.59 83.00 (77.50; 89.50)	0.694	0.787	79.32 ± 13.02 79.00 (71.00; 86.00)	82.78 ± 12.10 83.00 (77.00; 90.00)	79.62 ± 11.99 80.00 (70.00; 86.00)	0.174	0.053
Max heart rate		158.00 ± 19.08 159.00 (146.50; 169.50)	161.51 ± 16.75 163.00 (152.50; 171.00)	0.321	0.955		153.08 ± 18.28 155.00 (137.00; 171.00)	155.65 ± 17.70 157.00 (142.00; 168.00)	0.354	0.013
VO2 peak	–	22.05 ± 5.47 21.65 (18.03; 25.00)	24.15 ± 5.57 24.03 (20.82; 27.52)	–	< 0.001	–	20.68 ± 4.32 20.66 (17.58; 23.51)	21.45 ± 4.32 21.25 (17.51; 24.98)	–	0.230
VE peak	–	61.07 ± 17.01 58.65 (49.62; 70.34)	62.59 ± 16.32 60.00 (51.89; 72.22)	–	0.154	–	59.40 ± 12.60 59.90 (49.66; 64.71)	59.41 ± 12.14 57.10 (52.07; 67.12)	–	0.774

Continuous variables are expressed as mean ± standard deviation and as median with interquartile range. Differences between times points are represented by p values obtained using the Wilcoxon Signed-Rank Test.

Both groups showed a significant improvement in mean walking distance and estimated peak oxygen uptake (eVO2peak) during the recovery phase, however, the intergroup comparison did not show statistically significant difference in the improvement of monitored parameter between the groups (Table 3).

Table 3.

Change of selected parameters after exercise or recovery, stratified by group.

Variable	Change	Experimental group (n = 35)	Control group (n = 37)	p value
6-min walk distance	Absolute	32.3 ± 37.8 25.0 (10.8; 52.8)	26.2 ± 38.3 24.5 (− 7.5; 47.0)	0.395
	Percent	5.9 ± 6.9 5.0 (1.9; 9.4)	5.8 ± 9.3 4.4 (− 1.5; 9.8)	0.499
eVO2 peak	Absolute	0.7 ± 0.9 0.6 (0.2; 1.2)	0.6 ± 0.9 0.6 (− 0.1; 1.1)	0.460
	Percent	4.3 ± 4.9 3.5 (1.4; 6.8)	4.0 ± 5.8 3.2 (− 0.8; 6.8)	0.558
VO2 peak	Absolute	2.1 ± 2.5 2.5 (0.7; 3.7)	0.8 ± 2.5 − 0.0 (− 0.6; 2.0)	0.009
	Percent	10.4 ± 11.1 11.0 (2.9; 16.2)	4.6 ± 13.0 − 0.1 (− 2.6; 11.8)	0.018
VEpeak	Absolute	1.5 ± 9.4 2.6 (− 2.8; 7.4)	0.0 ± 6.8 − 1.1 (− 2.8; 5.4)	0.300
	Percent	3.9 ± 16.0 4.0 (− 5.1; 12.5)	0.8 ± 11.5 − 1.2 (− 4.7; 9.3)	0.308

Both absolute and relative changes are expressed as mean ± standard deviation and as median with interquartile range. Differences between the experimental and control groups are represented by p values obtained using the Mann–Whitney U test.

Contrarily, when measured by cardiopulmonary exercise testing (CPET), the mean VO2peak in the SAPA group increased significantly after the intervention (2.1 ml kg⁻¹ min⁻¹; p < 0.001) compared to only a slight increase in the control group (0.77 ml kg⁻¹ min⁻¹; p = 0.230), which was also reflected in the significant difference in intergroup comparison (p = 0.009), as presented in Table 3. Peak minute ventilation did not show significant group changes after the 12-weeks interval post chemotherapy.

Heart rate variability

Table 4 outlines longitudinal changes in heart rate variability (HRV) parameters throughout the study, with the intercepts indicating mean baseline values.

Table 4.

Linear mixed-effect model of HRV parameter changes (N = 72).

Variable		Estimate	Standard Error	T-value	p value
TotPwr	Intercept	1358.05	127.01	10.69	–
	Time since chemotherapy (months)	− 14.00	18.04	− 0.78	0.439
	Time since exercise or recovery (months)	122.84	38.06	3.23	0.001
	Exercise effect (months)	63.67	45.36	1.40	0.161
PLO	Intercept	421.46	43.10	9.78	–
	Time since chemotherapy (months)	− 2.89	7.41	− 0.39	0.697
	Time since exercise or recovery (months)	36.16	16.73	2.16	0.031
	Exercise effect (months)	26.39	19.25	1.37	0.171
PHI	Intercept	557.83	73.11	7.63	–
	Time since chemotherapy (months)	− 3.88	8.99	− 0.43	0.666
	Time since exercise or recovery (months)	43.82	20.33	2.16	0.031
	Exercise effect (months)	4.68	23.83	0.20	0.844
Ratio LO/HI	Intercept	1.61	0.16	10.10	–
	Time since chemotherapy (months)	− 0.03	0.02	− 1.26	0.209
	Time since exercise or recovery (months)	− 0.07	0.05	− 1.55	0.121
	Exercise effect (months)	0.06	0.05	1.14	0.256
MSSD	Intercept	1714.19	216.27	7.93	–
	Time since chemotherapy (months)	0.92	33.42	0.03	0.978
	Time since exercise or recovery (months)	169.00	67.96	2.49	0.013
	Exercise effect (months)	27.02	82.12	0.33	0.742

p values were obtained from linear mixed-effects models using Satterthwaite’s approximation for degrees of freedom.

During the chemotherapy period, no significant time-dependent changes were observed for neither parameter (all p > 0.20), although the slopes were negative accross most HRV outcomes, pronounced mainly in total power (TotPwr) (− 14.00 units/month, p = 0.439). In contrast, during the post-chemotherapy period, we found a consistent improvement in HRV marked by significant monthly increases detected in TotPwr (estimate =  + 122.84 units/month, p = 0.001), LF component—PLO (estimate =  + 36.16 units/month, p = 0.031), HF component—PHI (estimate =  + 43.82 units/month, p = 0.031) as well as MSSD (estimate =  + 169.00 units/month, p = 0.013). No significant recovery-related change was observed for the Ratio LO/HI (p = 0.121), which reflects parallel LF and HF increases with stable ratio.

The additional effect of the exercise intervention, modeled as an incremental monthly change during the post-chemotherapy period, was not statistically significant for any HRV parameter (all p > 0.15).

Changes in body composition

At baseline, the groups were balanced for all monitored variables, as shown in Table 5, with the p value for baseline (T0) intergroup comparison of 0.400 for BMI and 0.629 for the other monitored variables.

Table 5.

Comparison of body composition parameters before and after chemotherapy and intervention, stratified by group.

Variable	Experimental group (n = 35)					Control group (n = 37)					Experimental versus control group
	Baseline (T0)	After chemotherapy (T1)	After final assessments (T2)	p value (T0 vs T1)	p value (T1 vs T2)	Baseline (T0)	After chemotherapy (T1)	After final assessments (T2)	p value (T0 vs T1)	p value (T1 vs T2)	p value (T0)	p value (T1)	p value (T2)
Total Body Water	35.19 ± 4.29 34.40 (32.10; 38.25)	36.13 ± 4.21 35.50 (33.15; 39.55)	35.55 ± 3.78 35.00 (32.45; 38.45)	0.003	0.028	34.93 ± 4.24 34.25 (31.48; 37.07)	36.94 ± 4.75 35.10 (33.68; 40.02)	35.99 ± 4.43 35.05 (33.10; 38.95)	< 0.001	0.001	0.629	0.629	0.629
Percent Body Fat	32.09 ± 8.45 33.50 (27.05; 37.70)	30.79 ± 8.56 30.90 (25.00; 37.20)	31.55 ± 8.37 32.90 (25.65; 37.85)	0.110	0.182	34.63 ± 8.08 35.60 (30.05; 39.82)	32.91 ± 7.85 34.00 (29.43; 37.95)	33.83 ± 8.78 35.85 (28.95; 39.73)	0.159	0.024	0.629	0.629	0.629
Lean Body Mass	47.88 ± 5.85 46.90 (43.65; 52.25)	49.23 ± 5.68 48.40 (45.15; 53.80)	48.42 ± 5.17 47.90 (44.30; 52.40)	0.001	0.027	47.62 ± 5.77 46.70 (43.00; 50.48)	50.29 ± 6.37 47.85 (45.88; 54.60)	49.12 ± 6.11 47.85 (45.20; 52.95)	< 0.001	0.003	0.629	0.629	0.629
Weight	71.98 ± 15.38 71.20 (61.00; 76.60)	72.49 ± 14.65 70.80 (61.00; 78.55)	72.11 ± 14.04 72.30 (61.10; 78.00)	0.242	0.166	73.28 ± 11.98 68.90 (66.60; 81.50)	75.87 ± 12.71 74.45 (65.95; 82.67)	75.18 ± 12.08 74.20 (65.93; 83.70)	0.004	0.203	0.629	0.400	0.629
BMI	25.62 ± 5.24 24.81 (21.80; 28.23)	25.89 ± 5.02 25.90 (21.95; 29.18)	25.78 ± 5.02 25.71 (22.04; 28.80)	0.124	0.222	26.16 ± 4.67 25.08 (23.11; 28.27)	27.07 ± 4.61 26.01 (23.47; 29.74)	26.87 ± 4.49 26.12 (23.57; 29.26)	0.003	0.192	0.400	0.629	0.857

Continuous variables are expressed as mean ± standard deviation and as median with interquartile range. Differences between times points are represented by p values obtained using the Wilcoxon Signed-Rank Test. Differences between the experimental and control groups are represented by p values obtained using the Mann–Whitney U test.

Looking at the effect of chemotherapy on the anthropometry parameters prior to the intervention, it increased BMI while maintaining favorable changes in body fat and lean body mass in both groups, reaching statistical significance mostly in the control group.

After the intervention, we noted a significant decrease in lean body mass and total body water in both groups, though there was no statistically significant difference noted in intergroup comparison (Table 5).

Patient reported outcomes

There were no differences in patient reported scores for fatigue, physical functioning and global quality of life noted between the groups at the baseline. Both groups reported significant deterioration of physical functioning after the completion of chemotherapy (p < 0.001), whereas worsening of fatigue and global QoL was not statistically significant. The analyses revealed a significant improvement of physical and global scores in the SAPA group at 12 weeks after chemotherapy completion (p = 0.030 and 0.046 resp) in contrast to control group, where the differences were less pronounced. Both groups reported only a minor improvement in fatigue. The intergroup comparisons, however, did not reach statistical significance. Figure 3 illustrates the comparison of relative changes in all monitored dimensions of QoL between the groups.

Fig. 3.

Comparison of changes in health-related quality of life between groups.

Discussion

This study investigated the effects of a supervised, online-delivered exercise intervention on cardiorespiratory fitness, autonomic nervous system function, body composition, and quality of life in breast cancer survivors undergoing adjuvant chemotherapy. Our findings indicate that while the intervention significantly improved peak aerobic capacity measured by cardiopulmonary exercise testing, it did not lead to additional improvements in heart rate variability beyond those observed during the natural post-chemotherapy recovery period. Changes in body composition and patient-reported outcomes did not significantly differ between the intervention and control groups.

The most robust intervention-related effect was observed in peak oxygen consumption. Participants in the SAPA group achieved a significant increase in VO₂peak, exceeding the modest improvement seen in the control group. This finding is consistent with existing evidence indicating that aerobic training prescribed relative to individual cardiorespiratory capacity (typically within 60–80% of VO₂peak) is effective in improving aerobic fitness among cancer survivors^{17,22–25,27}.

In contrast, functional capacity assessed by the 6-min walk test and derived eVO2peak did not differ significantly between groups and neither did it show marked deterioration following chemotherapy. As the test was originally developed for patients with significant cardiovascular impairment, it may be less sensitive and reliable for detecting subtle changes in cardiorespiratory function among breast cancer survivors. This observation aligns with previous studies that reported only moderate or poor concordance between 6-min walk test–derived estimates and directly measured CPET parameters in oncology populations, particularly during or shortly after systemic treatment^28,29.

Contrary to expectations and some previous reports^30–34, the exercise intervention did not exert an independent effect on HRV outcomes. Longitudinal mixed-effect modelling revealed a consistent and significant improvement of HRV parameters—including total power, low- and high-frequency components, and MSSD—during the post-chemotherapy period, irrespective of group allocation. Importantly, the observed parallel recovery of low- and high-frequency HRV components with a stable LO/HI ratio suggests a generalized restoration of autonomic flexibility rather than a selective shift in sympatho-vagal balance.

Several factors may explain the absence of an intervention-specific improvement of HRV parameters. First, although the prescribed exercise intensity and adherence to the training was theoretically sufficient to elicit vagal and sympathetic adaptations, compliance with repeated home-based HRV monitoring declined toward the end of the intervention period. While linear mixed-effects modelling partially compensates for unequal numbers of observations, reduced data density may have limited the statistical power to detect small exercise-related changes in HRV. Second, short-term HRV measurements, particularly when conducted outside laboratory environments, are highly sensitive to situational influences, such as psychological stress, dietary habits, and environmental factors^35–37. Although detailed standardized instructions were provided to participants in order to minimize contextual confounders, these uncontrolled factors may have increased measurement variability and diluted intervention-specific effects.

Moreover, frequency-domain HRV parameters are known to exhibit substantial contextual variability, particularly in a clinically sensitive stage following breast cancer diagnosis, which should be considered when interpreting longitudinal HRV trajectories³⁸.

Thirdly, since physical activity outside the intervention was not quantified, it is also plausible that unstructured lower-intensity activities—such as household or other habitual physical activities not captured by exercise logs—contributed to the observed recovery in both groups. A substantial body of evidence indicates that light-intensity physical activity and reductions in sedentary behavior may exert independent and clinically meaningful effects on cardiovascular and autonomic health in cancer survivors^22,39,40. These forms of activity may be more sustainable during recovery phases and could partially explain the lack of differentiation between structured exercise and usual care in autonomic outcomes observed in this study.

Lastly, we acknowledge that the study was not powered to robustly detect differences in outcomes with high inter-individual variability, such as HRV indices. Therefore, non-significant findings may reflect limited statistical power rather than the lack of group differences. Our results should thus be interpreted with caution and viewed in the context of overall trends rather than definitive absence of group differences.

Neither body composition nor patient-reported outcomes demonstrated significant intergroup differences following the intervention. The most prominent chemotherapy-related anthropometric change was an increase in total body water, likely attributable to taxane-associated fluid retention⁴¹. This effect may have skewed lean body mass estimates derived from bioimpedance analysis, a method known to be sensitive to hydration status and physiological fluctuations in oncology populations⁴², potentially masking true changes in skeletal muscle mass⁴³. The subsequent decline in lean body mass during the recovery or intervention period should therefore be interpreted cautiously, as it may reflect the limitations of bioimpedance method rather than true muscular atrophy. The modest and comparable changes in BMI and body fat observed in both groups may also reflect the relatively short intervention duration, the influence of adjuvant endocrine therapy, and unmeasured lifestyle factors such as dietary intake. In this context, concurrent assessment of exercise-induced bioactive molecules could provide a more sensitive and mechanistically informative complement for monitoring training-related metabolic signaling that underlies changes in fat and lean body mass.

Improvements in physical functioning and global quality of life within the SAPA group, although clinically relevant, did not translate into statistically significant intergroup differences. This pattern suggests that gains in cardiorespiratory fitness may represent a relatively proximal physiological adaptation that does not automatically translate into broader quality-of-life benefits. Moreover, the extent to which patients can convert this physiological reserve into meaningful changes in QoL domains may be strongly influenced by other confounders such as ongoing treatment effects, psychosocial stressors, and lifestyle factors. The absence of significant between-group QoL differences may also be attributable to the delivery format of the intervention that lacked some key elements associated with greater quality-of-life improvements, such as in-person supervision, peer interaction, and group cohesion^44,45. These mechanisms can not only support training adherence and aerobic improvements, but may also be relevant for autonomic adaptations (e.g., affective/social engagement, recovery behaviors, and the training dose actually achieved).

Strengths of this study include the individualized exercise prescription based on objective cardiopulmonary testing, longitudinal monitoring of HRV and subsequent analysis using mixed-effects modelling. Limitations include a relatively high attrition rate, reduced adherence toward the end of the intervention, and limited control over confounding lifestyle and environmental factors affecting HRV measurements. This study’s findings should also be interpreted in light of its single-center design and relatively homogeneous sample, which may limit external validity. As a result, the observed effects may not fully generalize to broader and more diverse survivorship populations.

In conclusion, our results indicate that an online, structured exercise program improved aerobic fitness after chemotherapy; however, we did not observe clear intergroup differences in HRV, body composition, or quality-of-life outcomes. Given the limited sample size, the study may have been underpowered to detect modest but clinically meaningful changes in these secondary endpoints. This highlights the need for strategies that strengthen recruitment and support sustained participation in structured exercise, including tailored motivational and behavioral approaches that promote long-term adherence. Future multi-center studies with more heterogeneous cohorts are therefore warranted to confirm generalizability and strengthen the evidence base for remotely supervised exercise after chemotherapy.

Methods

The patients were recruited and treated at the Masaryk Memorial Cancer Institute (Brno, Czech Republic) between October 2021 and June 2024. The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of Masaryk Memorial Cancer Institute (2020/1530/MOU). Written informed consent was obtained from all participants. Eligible patients included women indicated to adjuvant chemotherapy based on platinum compounds, taxanes or vinca alkaloids with ECOG performance status less than 2 and capable of walking at least 400 m without rest. Patients were excluded if they had metastatic cancer, uncontrolled disease of the lungs, joints or cardiovascular system, autoimmune disease, or they were treated with beta-blockers. Participants were randomized using a REDCap module either to a supported adapted physical activity (SAPA) or control (CO) group, stratified by age and ECOG parameters.

Experimental design

Prior to the initiation of adjuvant chemotherapy (T0), baseline measurements of endpoint parameters (body composition, 6-min walk test, quality of life questionnaires) were conducted. These parameters were reevaluated after the completion of adjuvant chemotherapy (T1). At this point, cardiopulmonary exercise test was performed to establish peak oxygen consumption (VO₂peak) as well as resting and maximum heart rate (HR). Based on these values, exercise intensity was individualized with a HR interval corresponding to 60–80% of VO₂peak for the SAPA group. Final assessments (T2) of all study endpoints were performed following a 12-week period of either exercise intervention program or usual care in all study participants. HRV monitoring was conducted continuously throughout the study (between T0 and T2).

Exercise intervention program

All participants were instructed on HR monitoring using mySASY heart rate monitor and the PolarBeat mobile application. Women assigned to the SAPA group underwent a 12-week home-based exercise intervention program conducted online via the ZOOM platform three times a week and led by a professional physical therapist.

Each session included a 10-min warm-up phase, 45-min combined aerobic-resistance interval training, followed by 10 min of static stretching, calming breathing exercises, and relaxation techniques. Participants were periodically prompted to verify that their HR remained within the prescribed range of 60–80% of their maximum oxygen consumption (VO2max). Exercise adherence was defined as the proportion of completed sessions performed within the predefined target heart rate zone, evaluated using the average heart rate recorded during each session. Perceived exertion was assessed after each session using the Borg RPE scale to provide additional context regarding exercise intensity. Women in the control group did not receive specific guidance for exercise and were instructed to continue their lifestyle habits as usual. After 12 weeks period they were invited to join the online-based exercise program.

Assessment of study endpoints

Body composition was assessed using the bioimpedance method with the InBody 230 device. The assessed parameters included Body Weight (W), Body Mass Index (BMI), Percentage of Body Fat (PBF), Lean Body Mass (LM) and Total Body Water (TBW).

In 6-Minute Walking Test, participants were required to walk as fast as possible for 6 min along a straight corridor around two cones 30 m apart. The distance traversed by each participant over the designated period was recorded as 6-min walked distance (6MWD). An additional outcome was the estimated peak oxygen uptake (eVO₂peak), calculated from the six-minute walk distance according to the equation proposed by Ross et al. (2010).

Cardiopulmonary exercise testing (CPET) was conducted and evaluated by a sports medicine physician. The MetaLyzer 3B-R3 and Lode-Excalibur 38k4 devices were utilized in the study. The initial load was set to 1W/ kg and will gradually increase by 0.3W/kg to the subjective or objective maximum. VO₂peak, maximum heart rate (HR), and resting HR were measured.

Heart rate variability was measured using mySASY® monitor and software (Olomouc, Czech Republic). Measurements were taken at home three times a week after waking up, without any distractions. To support adherence, both groups received a monthly reminder via SMS. The measurement protocol involved an orthostatic test (a supine phase with 120 HR beats, followed by a standing phase with 360 HR beats, and concluding with supine phase with 360 HR beats). The whole protocol took about 15 min per session. Selected data were further evaluated using spectral HRV analysis, yielding total power (TotPwr), low frequency band power (PLO), high frequency band power (PHI), LO/HI ratio and MSSD values.

Patient reported outcomes were measured by the European Organization for Research and Treatment of Cancer Core Cancer Quality of Life Questionnaire (EORTC QLQ-C30).

Statistical analysis

Comparisons of baseline patient characteristics between the SAPA and CO groups were summarized with counts and frequencies and tested by Fisher’s exact test. Continuous variables were expressed as mean ± standard deviation and median (interquartile range) and compared using the Mann–Whitney U test. The evaluation of changes in parameters over the given period for the SAPA and CO group was performed using the Wilcoxon signed-rank test. The comparison of parameter changes between groups was tested using the Mann–Whitney U test.

Linear mixed-effects models were used to examine longitudinal changes in HRV parameters derived from the third supine phase of the orthostatic test. At baseline, mean values were estimated by the model intercepts.

Author contributions

Conceptualization, JH, LB; Data curation, JH, LB, KŠ; Formal analysis, JH, LB, KŠ; Funding acquisition, JH; Investigation, JH, LB, MP, KP, IH, LG, IB, PB; Methodology, JH, KŠ, LB, IB; Project administration, JH; Writing—original draft, LB, KŠ, JH; Writing—review & editing, JH, LB, IK, MP, KP, IH, LG, MŘ, IB, KŠ, EŘ. Supervision JH.

Funding

This research was funded by the Ministry of Health of the Czech Republic, MZ ČR—RVO (MOÚ, 00209805), and grant NU21-09-00558. All rights reserved. The trial was registered in the National database of scientific projects (IS VAVAI) under the registration number NU21-09-00558 on 1.5.2021, available at: (https://www.isvavai.cz/cep?s=rozsirene-vyhledavani&ss=detail&n=0&h=NU21-09-00558.)

Data availability

The datasets generated and analyzed in the current study are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.