With the intensification of population aging in China, the issue of frailty among the elderly is becoming increasingly prominent, making research on its prevention and intervention particularly important. Currently, most studies lack discussion on the dynamic relationship between changes in physical activity and frailty.

This study is based on the five waves of data from the China Health and Retirement Longitudinal Study (CHARLS) from 2011 to 2020, aiming to explore the association between physical activity changes trajectories and frailty in older adults, and to provide a scientific basis for the prevention and intervention of frailty in the elderly.

Group-based trajectory modeling (GBTM) was used to identify the potential subgroups and trajectory characteristics of physical activity over time among the survey participants during the follow-up period. Multivariate unconditional logistic regression models were employed to analyze the association between different physical activity trajectory types and frailty, as well as subgroup analyses.

The physical activity trajectories of the survey participants were divided into four groups: persistent low group (262 individuals, 13.87%), low-to-increasing group (993 individuals, 52.57%), high-to-decreasing group (122 individuals, 6.46%), and persistent high group (512 individuals, 27.10%). There were significant differences in frailty among the four groups (χ²=20.867, P<0.001). After adjusting for confounding factors such as age and gender, multivariate unconditional Logistic regression indicated that compared with the persistent low group, the low-to-increasing group (OR=0.581, 95%CI=0.414-0.815, P=0.002) and the persistent high group (OR=0.546, 95%CI=0.373-0.799, P=0.002) had significantly lower risks of frailty. Subgroup analysis revealed that, compared with the consistently low group, the initially low then rising group demonstrated significant reductions in frailty risk among the following elderly subgroups: age≥65 years (OR=0.502, 95%CI=0.345-0.730), males (OR=0.539, 95%CI=0.326-0.891), urban residents (OR=0.441, 95%CI=0.211-0.922), those without a partner (OR=0.312, 95%CI=0.160-0.606) (P<0.05). Similarly, the consistently high group exhibited protective effects against frailty in elderly individuals aged≥65 years (OR=0.425, 95%CI=0.274-0.658), females (OR=0.539, 95%CI=0.328-0.886), urban residents (OR=0.280, 95%CI=0.101-0.780), and those without a partner (OR=0.347, 95%CI=0.164-0.737) (P<0.05).

Different trajectory groups are associated with the risk of frailty. Physical activity trajectories characterized by a low-to-increasing pattern and persistent high levels can significantly reduce the incidence of frailty in older adults.

随着我国人口老龄化加剧，老年人的衰弱问题日益突出，对其预防和干预研究显得尤为重要。目前，大部分研究缺乏身体活动变化轨迹与衰弱动态关系的探讨。

探究老年人身体活动变化轨迹与衰弱的关联，为老年人衰弱的预防和干预提供科学依据。

本研究基于2011—2020年中国健康与养老追踪调查（CHARLS）5期数据，采用组轨迹模型（GBTM）识别调查对象随访期间身体活动随时间变化的潜在分组和轨迹特征。通过多因素Logistic回归模型分析不同身体活动轨迹类型与衰弱的关联及亚组分析。

共纳入1 889名老年人，其中男1 014名（53.7%）、女875名（46.3%），平均年龄（68.76±6.31）岁，衰弱者318名（16.8%）。身体活动轨迹分为持续低组262名（13.87%）、先低后上升组993名（52.57%）、先高后下降组122名（6.46%）和持续高组512名（27.10%）；4组老年人衰弱状况比较，差异有统计学意义（χ²=20.867，P<0.001）。调整了年龄、性别等混杂因素后，多因素Logistic回归分析结果显示，与持续低组相比，先低后上升组（OR=0.581，95%CI=0.414~0.815，P=0.002）和持续高组（OR=0.546，95%CI=0.373~0.799，P=0.002）的老年人衰弱发生风险显著降低。亚组分析结果显示，与持续低组相比，先低后上升组可降低年龄≥65岁（OR=0.502，95%CI=0.345~0.730）、男性（OR=0.539，95%CI=0.326~0.891）、居住地为城镇（OR=0.441，95%CI=0.211~0.922）、无伴侣（OR=0.312，95%CI=0.160~0.606）老年人的衰弱风险（P<0.05），持续高组可降低年龄≥65岁（OR=0.425，95%CI=0.274~0.658）、女性（OR=0.539，95%CI=0.328~0.886）、居住地为城镇（OR=0.280，95%CI=0.101~0.780）、无伴侣（OR=0.347，95%CI=0.164~0.737）老年人的衰弱风险（P<0.05）。

不同身体活动轨迹组别与衰弱发生风险有关，身体活动轨迹先低后上升组和持续高组可显著降低老年人衰弱发生风险。