1.
Dear Editor‐in‐Chief,
Prof. Ji‐Guang Wang, Colleagues Zhu et al. [1] is worth an acknowledgment for their contribution in providing information about the association among biological age (BioAge‐gap), other ages (HeartAge‐gap), and incident stroke in a cohort of Chinese subjects (n = 7283). In one example, a statistically significant increase in the risk of having a stroke was observed when there was a higher BioAge‐gap (OR = 1.05) as well as a higher HeartAge‐gap (OR = 2.41) in the extremely high category. There were some parts of the methodology that could have contributed more clinically to understanding the topic better.
Firstly, although the outcome, new onset stroke, was defined by physician diagnosis, which is always preferred when measuring such outcomes, it was based solely on participant self‐report, an issue particularly disturbing in light of the high burden of stroke in China (prevalence = 2.6, incidence = 505.2/100 000 person years, mortality = 343.4/100 000 person years, resulting in 3.4 million incident strokes and 2.3 million deaths yearly), for which hypertension has been reported as prevalent even in this high‐risk population [2]. Such definitions pose the danger of recall bias among the less‐educated groups.
Secondly, BioAge‐gap and HeartAge‐gap were created, raising the potential problems of circularity and/or over‐adjustment, since these indices are constructed from traditional CV and metabolic biomarkers (e.g., SBP, TC, and CRP), which are used in turn as additional covariates in the multiple regressions. Even though sensitivity analyses have addressed this problem, it may very well constitute a rehashing of established associations rather than an attempt to identify truly independent associations between the two constructs being measured.
Thirdly, there is little reported difference in predictive improvement (AUC difference from 0.664 to 0.677; NRI = 0.204). Despite being statistically significant, such differences may not be meaningful, since even small changes in discriminatory statistics could overestimate the added value of novel predictors without improving risk stratification or medical decisions [3].
Lastly, no time‐to‐event analyses were conducted. Without valid time‐varying information about stroke incidences, it is impossible to apply logistic regression methods to incorporate differing follow‐up periods or competing events, such as mortality [4].
In summary, it is necessary to emphasize that the research by Zhu et al. has made an important contribution to the issue under discussion. However, the above‐mentioned methodological aspects must be considered. It is critical to carry out future research that would use adjudicated events and time‐to‐event analyses.
Author Contributions
The author solely contributed to the planning, writing, and revision of the manuscript.
Funding
No funding was received for this study.
Consent
The author has nothing to report.
Conflicts of Interest
The author declares no conflicts of interest.
Data Availability Statement
Data sharing does not apply to this article as no datasets were generated or analyzed during the current study.
References
- 1. Zhu H., Lui H., He L., Zhang J., and Jin Z., “Associations of Biological Age and Heart Age Accelerations With New‐Onset Stroke in Individuals With Cardiovascular‐Kidney‐Metabolic Syndrome Stages 0–3: Evidence From a Chinese Longitudinal Study,” Journal Of Clinical Hypertension 28, no. 4 (2026): e70256, 10.1111/jch.70256. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Tu W. J., Zhao Z., Yin P., et al., “Estimated Burden of Stroke in China in 2020,” JAMA Network Open 6, no. 3 (2023): e231455, 10.1001/jamanetworkopen.2023.1455. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Pencina M. J., R. B. D' Agostino, Sr , R. B. D' Agostino, Jr , and Vasan R. S., “Evaluating the Added Predictive Ability of a New Marker: From Area Under the ROC Curve to Reclassification and Beyond,” Statistics in Medicine 27, no. 2 (2008): 157–172, 10.1002/sim.2929. [DOI] [PubMed] [Google Scholar]
- 4. Fine J. P. and Gray R. J., “A Proportional Hazards Model for the Subdistribution of a Competing Risk,” Journal of the American Statistical Association 94, no. 446 (1999): 496–509, 10.1080/01621459.1999.10474144. [DOI] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data sharing does not apply to this article as no datasets were generated or analyzed during the current study.