Abstract: Background　Decreased exercise endurance still exists after atrial fibrillation radiofrequency ablation. A number of studies have proved that atrial fibrillation exercise rehabilitation is effective，and adherence is a key determinant of whether the patient continues to benefit from exercise therapy. Therefore，understanding the dynamic changes and influencing factors of patients' exercise adherence after atrial fibrillation radiofrequency ablation is of great significance to establish the mechanism of internalization of external dynamics in patients' home exercise rehabilitation. Objective　To analyze the longitudinal follow-up data of home-based exercise rehabilitation in patients with atrial fibrillation after radiofrequency ablation，reveal the dynamic change of exercise adherence and explore its influencing factors，in order to provide theoretical basis for the formulation of targeted intervention strategies. Methods　A total of 72 patients with atrial fibrillation after radiofrequency ablation who were selected by convenient sampling method from February 2019 to August 2020 in the First Affiliated Hospital of Nanjing Medical University，were treated with Internet+ home exercise rehabilitation intervention for 12 weeks，and followed up to 24 weeks. The average adherence rates in seven time periods of 0-1 week，2-4 weeks，5-8 weeks，9-12 weeks，13-16 weeks，17-20 weeks and 21-24 weeks were collected. A multi-stage latent growth curve model was constructed for exercise adherence，and robust maximum likelihood estimation（MLR）was used to estimate the mean，the variance and the covariance. Taking gender，education，working status，the existence of comorbidities and frequency adherence as covariates，the latent growth curve model was constructed for intensity-time adherence and the parameters were estimated by MLR method. Results　Intensity-time adherence and frequency adherence both showed a second-stage curve increasing trend model，suggesting that there is a turning point in the development of adherence，in the 12th week of exercise. The first-stage of intensity-time adherence（from the beginning to 12th week）showed a curve increasing trend，and the increasing rate slowed down with time（Q1=-0.062，P<0.05）；The second stage（from 12 weeks to 24 weeks of exercise）showed a linear downward trend（S2=-0.081，P<0.05），and the differences in the initial level and growth rate of individuals were statistically significant（P<0.05）. Frequency adherence showed a quadratic growth trend in the first stage，reaching the peak at the 4th week of exercise，and then decreased slightly；The second stage showed a quadratic downward trend，and the rate of decline accelerated with time（Q2=0.003，P<0.05），and the differences between the individual's initial level and the acceleration of change were statistically significant（P<0.05）. Among the time-constant covariates，only the existence of comorbidities had a significant impact on intensity-time adherence at the first week（P<0.05）. Between the first week and the 24th week，frequency adherence had a positive impact on intensity-time adherence（P<0.05）. Conclusions　During the intervention period，the level of intensive-time adherence and frequency adherence are relatively low，both with a curve increasing trend and stabilizing at a higher level，while the adherence decreased gradually during the follow-up period. Patients with comorbidities had poor initial intensity-time adherence，while frequency adherence had a positive impact on intensity-time adherence.

Key words: Atrial fibrillation, Internet, Home-based exercise, Cardiac rehabilitation, Radiofrequency ablation, Adherence, Development trajectory

摘要： 背景　心房颤动（以下简称房颤）射频消融术后患者运动耐力下降依然存在，多项研究证明房颤运动康复行之有效，依从性是决定患者运动治疗是否持续获益的关键环节。因此，了解房颤射频消融术后患者运动依从性的动态变化规律及影响因素对建立患者居家运动康复外部动力内化的机制具有重要意义。目的　分析房颤射频消融术后患者居家运动康复的纵向随访数据，揭示运动依从性的动态变化规律，并探讨其影响因素，为制定针对性干预策略提供理论依据。方法　采用便利抽样法，选取2019年2月至2020年8月就诊于南京医科大学第一附属医院心血管内科行射频消融手术的72例房颤患者为研究对象，对其进行互联网+居家运动康复干预12周，并随访追踪至24周。收集运动后0~1周、2~4周、5~8周、9~12周、13~16周、17~20周、21~24周7个时间段内平均依从率，利用多阶段潜变量增长模型对数据进行拟合，采用稳健最大似然估计（MLR）对均值、方差和协方差进行估计；以性别、教育程度、工作状态、是否存在合并症及频率依从性为协变量，对强度-时间依从性进行潜变量增长建模，采用MLR法对参数进行估计。结果　强度-时间依从性和频率依从性均为二阶二次型发展趋势模型，即依从性的发展存在转折点，在运动第12周。强度-时间依从性第一阶段（从开始到运动12周）呈二次型增长趋势，增长速度随时间减缓（Q1
=-0.062，P<0.05），第二阶段（从运动12周到24周）呈线性下降趋势（S2=-0.081，P<0.05），且个体在初始水平和变化速度上差异均有统计学意义（P<0.05）；频率依从性第一阶段呈二次型增长趋势，运动4周达到峰值，而后略有降低，第二阶段呈二次型下降趋势，下降速度随时间加快（Q2=0.003，P<0.05），且个体在初始水平和变化加速度上差异均有统计学意义（P<0.05）。在时间恒定协变量中，仅有是否存在合并症对患者运动1周内依从性水平存在影响（P<0.05）。运动1周至运动24周，患者运动频率依从性对强度-时间依从性均具有正向影响（P<0.05）。结论　运动干预期间，强度-时间依从性和频率依从性均逐渐升高并稳定于较高水平，随访期间依从性逐渐下降；存在合并症的患者初始强度-时间依从性水平较差，而频率依从性能正向预测强度-时间依从性的变化。

关键词: 心房颤动, 互联网, 居家运动, 心脏康复, 射频消融术, 依从性, 发展轨迹