Letter to the editor regarding: ‘global, regional, and national burden of chronic respiratory diseases, 1990–2021 and predictions to 2035: analysis of data from the global burden of disease study 2021’

Dear Editor

We read with great interest the study by Zhai et al. [1], which comprehensively analyses the evolving global and regional burden of chronic respiratory diseases (CRDs) over the past 30 years and their attributable risk factors. Although the study provides unique insights into the global trends of CRDs, its evaluation of temporal trends primarily relies on the estimated annual percentage change (EAPC) model, which assumes a linear trend over time. However, the actual changes in disease burden often follow a nonlinear or stage-specific pattern, suggesting that such linear assumptions may limit a comprehensive understanding of the true epidemiological dynamics. Therefore, caution is needed when interpreting the time trend results of the study, especially considering that disease burden changes may be influenced by multiple factors.

To investigate whether the global burden of CRDs during this period showed a linear or non-linear trend, we searched PubMed for all studies on CRDs trends. We found that Cao et al. [2] analysed CRDs trends over time using a different model, namely the joinpoint regression model. This model uses the average annual percentage change (AAPC) to calculate the geometric mean of the annual percentage changes over a specified time period, thereby obtaining an average annual change value that effectively mitigates the impact of inter-annual fluctuations [3].

The results of Cao et al.’s joint point regression analysis show that from 1990 to 2021, the AAPC for age-standardized incidence rate (ASIR), prevalence rate (ASPR), mortality rate (ASMR), and disability-adjusted life years (DALYs) rate (ASDR) of CRDs were −0.87% (95% confidence interval [CI]: −0.94% to −0.8%), −1.01% (95% CI: −1.1% to −0.94%), −1.46% (95% CI: −1.78% to −1.01%), and −1.51% (95% CI: −1.75% to −1.13%), respectively, with all indicators showing an overall downward trend. Notably, when we compare Cao et al.’s AAPC results with the EAPC reported by Zhai et al. in Table 1 of their study, we found only minor discrepancies between the two estimates. Zhai et al. reported the EAPC for ASIR, ASPR, ASMR, and ASDR of global CRDs from 1990 to 2021 as −0.79% (95% CI: −0.9% to −0.68%), −0.96% (95% CI: −1.06% to −0.87%), −1.68% (95% CI: −1.77% to −1.6%), and −1.68% (95% CI: −1.74% to −1.62%), respectively. Overall, although there are differences between AAPC and EAPC in quantifying time trends in CRDs, both methods show highly consistent overall trends. Specifically, over the past three decades, there has been a continuous downward trend in the global incidence, prevalence, mortality, and DALYs burden of CRDs.

The core advantage of the joinpoint regression model is its ability to identify points where trends change significantly and to account for fluctuations that simple percentage change calculations might obscure. In contrast, EAPC is more flexible and precise, capturing annual changes and short-term fluctuations, making it particularly suitable for data with significant annual variability. Based on this, we recommend using both AAPC and EAPC metrics in future epidemiological analyses to balance precision with interpretability. We are grateful for the opportunity to reflect on the methodology and trust that these insights will strengthen future assessments in global burden studies.

Funding Statement

This study was supported by a grant from the Westlake Laboratory of Life Sciences and Biomedicine (Grant No. XHSYS-02). The funding source had no role in the design, conduct, or analysis of this study.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data availability is not applicable to this article as no new data were created or analysed in this study.