Dynamic titration of exercise intensity in pulmonary rehabilitation

Pulmonary rehabilitation (PR) is a cornerstone of management for individuals with chronic obstructive pulmonary disease (COPD), improving exercise tolerance, dyspnea, functional capacity, and health-related quality of life. However, conventional exercise prescription frequently relies on baseline assessments and predetermined progression schedules that may not adequately account for day-to-day fluctuations in symptoms, oxygenation, cardiovascular responses, fatigue, and recovery. Dynamic titration of exercise intensity offers an alternative approach in which the exercise workload is adjusted according to the patient's physiological and perceptual responses before and during each rehabilitation session. Recent advances in artificial intelligence (AI), wearable monitoring, and machine learning provide opportunities to support this process by integrating modified Borg dyspnea scores (mBorg), resting peripheral oxygen saturation (SpO₂), heart rate, heart rate recovery, previous-session performance, and other relevant clinical variables into individualized exercise recommendations. This article examines the rationale, clinical applications, potential benefits, and implementation challenges of AIsupported dynamic exercise titration in COPD rehabilitation. Particular attention is given to three clinically meaningful outcomes: change in six-minute walk distance (6MWD), quadriceps muscle strength, and the frequency of exercise-induced oxygen desaturation. A physiotherapy-oriented framework is proposed in which AI estimates daily exercise readiness, recommends an initial walking speed or resistance-training load, and monitors the response to exercise, and supports subsequent progression or regression under professional supervision. The approach is compared conceptually with traditional static progression, while distinguishing established evidence for pulmonary rehabilitation and digital monitoring from the still-developing evidence for AIdirected daily exercise dosing. Safety, interpretability, patient autonomy, and clinical accountability remain essential. Prospective randomized controlled trials are required to establish whether AI-driven adaptive prescription produces superior functional outcomes and fewer desaturation events than well-delivered conventional pulmonary rehabilitation.