A Commentary on “Depressive Symptom Trajectories and Hypertension Risk Among Chinese Middle‐Aged and Older Adults”

We read with great interest the recent article by Sun et al., which examined the association between depressive symptom trajectories and hypertension risk among middle-aged and older adults in China 1. Their findings provide valuable insights into how dynamic psychological changes may influence cardiovascular outcomes. However, we would nevertheless like to offer several comments that may help further clarify their results. Firstly, Sun et al. identified five depressive symptom trajectories based on group-based trajectory modeling (GBTM). However, prior studies have shown that GBTM-derived classifications may be influenced by polynomial order, starting values, and other model assumptions 2, 3. As these trajectory groups served as the foundation for all of the authors' subsequent analyses, confirming the stability of these classifications under alternative model specifications is essential. If the trajectory groupings are unstable, any subsequent comparison of hypertension risk across these groups may also be unreliable. As such, conducting sensitivity analyses using different polynomial orders, starting values, or distributional assumptions would greatly strengthen confidence in the validity of the trajectory structure. Second, although blood pressure was objectively measured in the 2011, 2013, and 2015 waves of the China Health and Retirement Longitudinal Study (CHARLS), in the 2018 and 2020 waves hypertension was determined solely by self-report due to the absence of any blood pressure measurements. Prior research has further shown that self-reported hypertension may not always correspond with clinically measured values, particularly among older adults 4. Given that hypertension is the primary study outcome, discrepancies in reporting accuracy could introduce differential misclassification across different depressive symptom trajectories. From our perspective, this point is crucial, as individuals with varying psychological states may have different likelihoods of reporting chronic conditions. Thus, examining the concordance between self-reported and measured hypertension in earlier CHARLS waves would help to clarify the reliability of the outcome and its influence on the observed associations. Third, in their analysis, Sun et al. adjusted for numerous baseline covariates, but did not incorporate time-varying factors. However, physiological parameters, chronic disease status, autonomic activity, inflammatory markers, and lifestyle behaviors fluctuate over time and influence both depressive symptoms and hypertension through action on interconnected biological pathways 5. Moreover, specific time-varying factors, such as any changes in medication use (particularly antihypertensive and antidepressant medications), body mass index fluctuations, smoking cessation or initiation, and the onset of new chronic conditions throughout follow-up may confound the observed associations. Given that the CHARLS collected these data across multiple waves, incorporating these time-updated variables would facilitate a more nuanced understanding of the temporal dynamics between depressive trajectories and hypertension risk. Despite these considerations, we wish to reiterate that Sun et al.'s study provides valuable insights into the long-term relationship between psychological well-being and hypertension among older adults. Our comments are intended to support further refinement of the research in this area and to enhance interpretation of their important findings. R.K. conceived the idea for the Letter and drafted the manuscript. K.K. contributed to the interpretation of the literature and critically revised the manuscript. Both authors read and approved the final version. We thank Editage (www.editage.jp) for English language editing. The authors have nothing to report. The authors have nothing to report. The authors have nothing to report. The authors declare no conflicts of interest. Data sharing not applicable to this article as no datasets were generated or analysed during the current study.