Hybrid cardiac rehabilitation in patients with durable left ventricular assist devices: the REHAB-ASSIST study

Figures

• 

  Figure 1. Central illustration of the multicenter REHAB-ASSIST study. Study protocol, sample size, and main and secondary outcomes. Created with BioRender.com. For quantitative variables, the paired Student's t-test was used. For qualitative variables, McNemar's test was applied for paired data. *VT1: first ventilatory threshold. Alternative text: Schematic overview of a multicenter proof-of-concept study including 15 patients with durable left ventricular assist devices recruited from two tertiary hospitals in Spain. The intervention consisted of a 12-week hybrid cardiac rehabilitation program delivered six days per week, combining three supervised center-based sessions and three home-based sessions via telemedicine using a mobile application and smartwatch. The program was multidisciplinary. Assessments were performed at baseline, at 3 months (end of program), and at 6 months (telemedicine follow-up). Clinical primary outcomes showed no significant change in peak oxygen consumption. Secondary clinical outcomes showed significant improvement in 6-minute walk distance and quality of life (Kansas City Questionnaire), significant improvement in ventilatory oscillations, and non-significant…

• 

  Figure 2. Study flow diagram. Created with BioRender.com. Alternative text: Flow diagram of patient selection and study participation. A total of 29 candidates with long-term ventricular assist devices were screened. Fourteen patients were excluded due to geographical barriers preventing participation in a center-based program. Fifteen patients were enrolled. Two patients withdrew: one due to relocation to another country before starting the program and one due to exacerbation of pre-existing pulmonary disease after a few sessions. Thirteen patients completed the hybrid cardiac rehabilitation program.

• 

  Figure 3. Box plots of peak oxygen consumption and 6-minute walk distance. For quantitative variables, the paired Student's t-test was used. Alternative text: Two sets of box plots comparing baseline, 3-month, and 6-month outcomes. The first set shows peak oxygen uptake (VO₂, mL/kg/min), with mean values of 11.2 at baseline, 12.2 at 3 months, and 12.4 at 6 months, without statistically significant differences. The second set shows 6-minute walk distance (meters), with mean values of 444 at baseline, 492 at 3 months, and 524 at 6 months, demonstrating a statistically significant improvement at follow-up. Each box plot displays median, interquartile range, and outliers.

• 

  Figure 4. Data from the telemedicine application and smartwatch. A generalized least squares (GLS) model with random effects was used to account for within-subject correlation. Alternative text: Two longitudinal line graphs showing daily step count and daily sleep duration over approximately 200 days of follow-up. Each graph includes a smoothed trend line with a shaded 95% confidence interval. Daily step counts fluctuate around approximately 4,000–5,000 steps per day without a statistically significant trend over time (random-effects generalized least squares model p = 0.135). Daily sleep duration averages approximately 4 to 6 h per night, also without a statistically significant trend (p = 0.154). The analyses account for within-subject correlation using a random-effects model.