We read with great interest the article by Liu Q et al. [1] examining the association between changes in physical activity and kidney function in the general population. The authors utilised a large cohort from the UK Biobank to investigate this important relationship, providing valuable insights. However, we would like to highlight a few key limitations and suggest future research directions that could strengthen the evidence in this area.

Firstly, the authors relied on self-reported physical activity data, which is subject to potential recall bias and social desirability bias. Individuals may overreport or underreport their physical activity levels, leading to misclassification and potentially biassing the observed associations [2]. The use of objective measures, such as accelerometers or activity trackers, could provide more accurate and reliable assessments of physical activity, reducing the risk of measurement error. Indeed, through UK Biobank, several studies have reported high quality research on accelerometer-measured physical activity and disease prognosis [3, 4].

Secondly, the authors' adjustment for covariates did not include dietary factors, which are known to have a significant impact on kidney function. Dietary intake of protein, sodium and other nutrients can influence serum creatinine and cystatin C levels, potentially confounding the relationship between physical activity and estimated glomerular filtration rate (eGFR) [5, 6]. Future studies should consider incorporating detailed dietary information, such as nutrient intake and dietary patterns, to better understand the complex interplay between physical activity, diet and kidney health.

Additionally, the authors focused their analysis on the general population, which may have different characteristics and risk profiles compared to individuals with pre-existing chronic kidney disease (CKD) or other comorbidities. It would be valuable to conduct subgroup analyses or stratified models to explore the potential differential effects of physical activity changes on kidney function in specific patient populations, such as those with CKD, diabetes or cardiovascular disease. This approach could provide more targeted insights and guide the development of tailored physical activity recommendations for individuals at higher risk of kidney dysfunction.

Furthermore, the authors utilised eGFR as the primary outcome, which is an estimated measure of kidney function. While eGFR is widely used in clinical practice, it may not accurately reflect true glomerular filtration rate, especially in the context of changing muscle mass and body composition associated with physical activity [7]. Future studies could consider incorporating direct measures of kidney function, such as iohexol or inulin clearance, to provide a more precise assessment of the relationship between physical activity and actual kidney function.

Finally, the authors' analysis was limited to two time points, which may not capture the dynamic nature of physical activity and its long-term impact on kidney health. Longitudinal studies with repeated assessments of physical activity and kidney function over an extended period could shed light on the trajectories of these variables and their interplay over time. This approach could help elucidate the causal mechanisms and identify critical time windows for interventions to preserve kidney function.

In conclusion, the authors' work provides valuable insights into the association between changes in physical activity and kidney function in the general population. However, addressing the limitations outlined above, such as the use of objective physical activity measures, incorporation of dietary factors, exploration of subgroup differences and the inclusion of direct kidney function assessments, could further strengthen the evidence and guide the development of targeted physical activity recommendations for kidney health.

我们饶有兴趣地阅读了 Liu Q 等人 [1] 的文章，该文章研究了普通人群身体活动变化与肾功能之间的关联。作者利用来自英国生物样本库的大量队列来调查这一重要关系，提供了有价值的见解。然而，我们想强调一些关键的局限性，并提出可以加强该领域证据的未来研究方向。

首先，作者依赖于自我报告的身体活动数据，这些数据受到潜在回忆偏倚和社会期望偏倚的影响。个体可能会高报或低报他们的身体活动水平，从而导致错误分类并可能使观察到的关联产生偏差 [2]。使用客观的测量方法，如加速度计或活动追踪器，可以提供更准确和可靠的身体活动评估，降低测量误差的风险。事实上，通过英国生物样本库，几项研究报告了关于加速度计测量的身体活动和疾病预后的高质量研究 [3， 4]。

其次，作者对协变量的调整不包括饮食因素，已知饮食因素对肾功能有显着影响。蛋白质、钠和其他营养素的膳食摄入会影响血清肌酐和胱抑素 C 水平，可能会混淆身体活动与估计肾小球滤过率 （eGFR） 之间的关系 [5， 6]。未来的研究应考虑纳入详细的饮食信息，例如营养摄入和饮食模式，以更好地了解身体活动、饮食和肾脏健康之间的复杂相互作用。

此外，作者将分析重点放在普通人群上，与患有慢性肾脏病 （CKD） 或其他合并症的个体相比，他们可能具有不同的特征和风险状况。进行亚组分析或分层模型以探索身体活动变化对特定患者群体（例如 CKD、糖尿病或心血管疾病患者）肾功能的潜在差异影响将是有价值的。这种方法可以提供更有针对性的见解，并指导为肾功能障碍风险较高的个体制定量身定制的身体活动建议。

此外，作者使用 eGFR 作为主要结局，这是肾功能的估计指标。虽然 eGFR 在临床实践中被广泛使用，但它可能无法准确反映真实的肾小球滤过率，尤其是在与体力活动相关的肌肉质量和身体成分发生变化的情况下 [7]。未来的研究可以考虑纳入肾功能的直接测量，例如碘海醇或菊粉清除率，以更精确地评估身体活动与实际肾功能之间的关系。

最后，作者的分析仅限于两个时间点，可能无法捕捉到身体活动的动态性质及其对肾脏健康的长期影响。对长时间的身体活动和肾功能进行重复评估的纵向研究可以阐明这些变量的轨迹以及它们随时间推移的相互作用。这种方法可以帮助阐明因果机制并确定保护肾功能的干预措施的关键时间窗口。

总之，作者的工作为普通人群身体活动变化与肾功能之间的关联提供了有价值的见解。然而，解决上述局限性，例如使用客观的身体活动措施、纳入饮食因素、探索亚组差异以及纳入直接肾功能评估，可以进一步加强证据并指导制定有针对性的身体活动建议以促进肾脏健康。