Depression, associated with low quality of life and increased health burdens, is a disabling and prevailing mental health issue (Sivertsen et al., 2015; Wilkinson et al., 2018), with more than 264 million people affected worldwide (James et al., 2018). Middle-aged and older adults living in resource-limited settings (e.g., China) are disproportionally affected by depression and physical disability (James et al., 2018). Although several theories have been used to interpret the potential association between depression and physical health, the inflammation process has drawn considerable attention from psychologists and psychiatrists (Raison et al., 2006; Miller et al., 2009; Slavich and Irwin, 2014). Several theories have interpreted the bidirectional association of depression and inflammation, and the macrophage theory of depression hypothesizes that the onset and progression of depression are associated with the amount of pro-inflammatory cytokines secreted by the macrophages (Smith, 1991). Inflammation plays an etiologic role in the development and progression of depression, probably by altering the neurotransmitter metabolism or hypothalamic–pituitary–adrenal (HPA) axis function (Raison et al., 2006). HPA hyperactivity and the autonomic nervous system have also been used to explain how depression affects the inflammation processes (Kop and Gottdiener, 2005). Previous meta-analytic studies reported that older participants with depression demonstrated a high level of dysregulation of HPA axis activity through several mechanisms, including physical illnesses (e.g., diabetes, cardiovascular disease, and autoimmune diseases), alterations in the central nervous system (CNS), and immune-endocrinological alterations. These proposed biological mechanisms suggest a bidirectional path of the depression-inflammation association. Additionally, inflammation may trigger all-out-effort sickness behaviors, many of which may be associated with depressive symptoms (Hart, 1988). At the same time, depression could be prospectively associated with inflammation via behavioral changes (such as exercise, diet, and substance use) that are associated with inflammation, simply increasing experienced stress (which might increase inflammation).

One of the commonly and frequently studied inflammation biomarkers is C-reactive protein (CRP). It is an acute phase plasma protein with a short plasma half-life and a relatively robust and reliable response to inflammation, making CRP an ideal marker of inflammation. The high-sensitivity test for CRP, called low-reactive protein (LRP, L-CRP, or hs-CRP), more accurately measures very low levels of CRP, and static sampling of CRP has been widely used in clinical studies to predict diseases, such as cancer and depression (Coventry et al., 2009; Felger et al., 2020; Nehring et al., 2022). In the development, progression and treatment of depression, an increasing number of researchers have focused on the association between CRP levels and depressive symptoms [including major depressive disorder (MDD) and treatment-resistant depression (TRD)] (Orsolini et al., 2022). Several meta-analyses have explored the association between CRP and depression in participants with either major or probable depression and found significant but small effect sizes (Howren et al., 2009; Valkanova et al., 2013; Haapakoski et al., 2015; Horn et al., 2018). More importantly, when considering the study quality of the original studies, the effect size was strikingly attenuated and nonsignificant in the most recent meta-analysis (Horn et al., 2018). Whether this association exists in studies conducted among middle-aged and older adults is highly debatable (Penninx et al., 2003; Almeida et al., 2007; Bremmer et al., 2008; Pan et al., 2008; Stewart et al., 2008; Surtees et al., 2008; Almeida et al., 2009; Elovainio et al., 2009; Milaneschi et al., 2009; Su et al., 2009; Forti et al., 2010; Slopen et al., 2010; Bjerkeset et al., 2011; Häfner et al., 2011; Morris et al., 2011; Baune et al., 2012; Smagula et al., 2014; Zoga et al., 2014; Song et al., 2015; Tully et al., 2015; de Menezes et al., 2017). Although most studies show a significant positive relationship in either depressed or nondepressed older adults, some well-conducted studies failed to replicate these results (Almeida et al., 2009; Su et al., 2009; Forti et al., 2010; Morris et al., 2011; Baune et al., 2012; Häfner et al., 2012; de Menezes et al., 2017). Most studies lack (1) a random sampling method, which may result in potential sampling bias; (2) longitudinal study design, which is needed for assessing the temporal relationship and bi-direction; (3) statistical strictness in choosing confounders, and (4) detailed report of data cleaning strategies.

The present study investigated the relationship between hs-CRP levels and depression scores in middle-aged and older adults using longitudinal analysis in a nationally representative cohort. We used available data from the China Health and Retirement Longitudinal Study (CHARLS), which was designed to describe and prospectively delineate the social-, economic-and functional status of middle-aged and older adults. The CHARLS database provides a unique opportunity to examine the temporal and bidirectional relationship between depression and hs-CRP in middle-aged and older adults over 5 years while controlling for various socio-economic, anthropometric, physiologic, and psychosocial factors.

This study investigated whether the inflammatory marker, CRP, could prospectively predict symptoms of depression in middle-aged and older adults using a prospective design of community and rural dwellings, accounting for a range of potential confounders. CRP was not significantly associated with depression both cross-sectionally and longitudinally in the largest sample. Consistent with the results of previous studies in some Asian regions, such as Islamabad, there was a lack of association between CRP and depression influenced by racial/ethnic, genetic, and environmental factors, and population-based assessments of associations between physiological processes or social integration should consider these variables (Chapman and Santos-Lozada, 2020; Zavos et al., 2022). Contrary to other reports, we did not find this association with gender modification either (Danner et al., 2003; Morris et al., 2011; Vogelzangs et al., 2012; Song et al., 2015; Köhler-Forsberg et al., 2017). In fact, it is unclear from existing studies whether CRP levels directly contribute to the onset and progression of depression.

Notably, BMI was negatively associated with depression in 2015 in our study, consistent with previous studies reported in Korea and China, and both studies demonstrated a negative correlation between BMI and depressive symptoms (Qin et al., 2017; Lee et al., 2019). This is contrary to the majority of findings from Western studies that reported a significant association between clinical overweighted or obesity and depression (Luppino et al., 2010; Daly, 2013; Preiss et al., 2013). Similarly, studies on other Asian populations have reported that being overweight or obese prevents depression. As previously reported, overweight individuals in Hong Kong and Bangladesh are less likely to have depressive symptoms, and obesity is an independent protective factor for depressive symptoms (Li et al., 2004). This counterintuitive epidemiological finding appears to be consistent with the “obesity paradox” theory, which reflects the relationship between obesity and reduced mortality compared with normal weight (Adams et al., 2006). However, the CESD-10 was used in the study to primarily reflect an individual’s experience of depression, and abnormalities in BMI (underweight and overweight) may affect a person’s body satisfaction and self-esteem, which in turn may affect the test results (Richard et al., 2016).

However, our study found no significant association between CRP level and depression both cross-sectionally and longitudinally. This is in contrast to most findings from Western studies, which reported a significantly higher prevalence of MDD in individuals with high CRP levels, and this correlation appears to be more prominent in younger adults than in older patients (Jung and Kang, 2019; Milaneschi et al., 2021; Orsolini et al., 2022). The cause of elevated CRP may be another potential mechanism of action for neurovascular injury induced by dysregulation of peripheral myeloid cells, pro-inflammatory cytokines and complement pathways (Menard et al., 2017; Horn et al., 2018). Additionally, elevated CRP levels and pro-inflammatory activity can drive inflammation through microglia and astrocyte activation (McKim et al., 2018; Wesselingh et al., 2019).

Our results lend partial support to studies on middle-aged and older adults that found no association between elevated CRP and higher odds of depression before and after adjusting for demographic, socio-economic, lifestyle, and prior medical history variables (Pan et al., 2008; Stewart et al., 2008, 2009; Song et al., 2015; de Menezes et al., 2017). Possible interpretations can be made based on two aspects: sampling methods and depression types. Some studies adopted a convenience sample, which may have caused selection bias. Our study adopted a four-stage random sampling method across China, minimizing the potential bias and magnifying the sample representativeness of older Chinese participants. Our results partially replicated studies conducted in some East Asian populations, who share similar lifestyles, such as a sample of 3,289 community residents from Beijing and Shanghai (Pan et al., 2008) and a sample of 569 from a rural community in Korea (Song et al., 2015). However, two studies that adopted a random sampling method, were conducted in Whites, and found a unidirectional association, which suggests that race may play a role in this association (Gimeno et al., 2009; Zalli et al., 2016). Owing to the highly mixed results in this field, ethnic and racial differences should also be considered in future studies. Third, CRP-depression may only exist in people living with atypical depression rather than melancholic depression (Glaus et al., 2014; Lamers et al., 2016), which reflects that depression subtypes may have unique but different biological pathways. Further studies are needed to verify the neurobiological pathways of the CRP-depression association in different subtypes.

The strengths and limitations of this study should be mentioned. The major power is the strictness of the study design (random sampling method, prospective cohort, and large sample size), which supports high internal and external validity. However, this study had several limitations. First, we did not collect other pro-inflammatory biomarkers, like interleukin-6 (IL-6), which might be longitudinally associated with future depression status (Stewart et al., 2009). Second, the CESD-10 is not a commonly used assessment of depression in previous reports, limiting the comparability of our results to those of other studies. Third, CRP was collected in two waves, which limits our ability to assess the dynamic interaction between CRP and depression. Fourth, the interval between participant visits is too long, which may cause dampening effects of potential associations or stronger effects due to long-term interactions between variables.

In summary, we failed to find a bidirectional association between CRP levels and depressive symptoms in well-represented middle-aged and older Chinese adults. Additionally, we did not observe any sex-related modifications. Future studies should validate the role of multifaceted pro-inflammatory biomarkers in predicting depression and should also consider racial/ethnic, genetic, and environmental factors for the lack of association between CRP and depression and the use of commonly used depression scales and smaller time intervals to make the results comparable.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary material.

Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

BC formulated the research questions, designed the study, supervised the data cleaning. XL wrote the paper. BC and YN carried out data consolidation, cleaning, and preliminary analysis. XL, BC, and YN were responsible for principal analysis and result interpretation. XL and BC were responsible for the revision of the manuscript. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.