With the global population continues to aged, the number of elderly individuals with mild cognitive impairment or dementia is dramatically rising in China and other countries (GBD 2016 Dementia Collaborators, 2019; Alzheimer’s Association, 2020; Jia et al., 2020). Mild cognitive impairment is considered to be the preclinical state of dementia and has gradually become an important global health problem, which affects people’s ability of learning and memory, reduces their quality of life, and has brought a heavy burden to their families and the whole country (Park et al., 2003; Liu et al., 2013). Thus, the early identification and prevention of cognitive impairment is of great significance in clinical practice.

Currently, more attention has been focused on searching blood biomarkers which is easy to obtain and cost-effective, since the positron emission tomography (PET) imaging is costly and lumbar puncture is unable to be generally accepted for diagnosing dementia (Feng et al., 2021). As we all know, serum creatinine and cystatin C are both biomarkers for glomerular filtration and renal function. Recently, a surrogate marker for sarcopenia called sarcopenia index (SI) has deen developed, which is defined as serum creatinine (mg/dL)/cystatin C (mg/L) × 100 and can be used to evaluate the skeletal muscle mass (Kashani et al., 2017; Barreto et al., 2019; Ren et al., 2022; Zheng et al., 2022). As we know, serum creatinine is the derivative of the skeletal muscle protein and the production of serum creatinine changes with skeletal muscle mass. In contrast, serum cystatin C is excreted by nucleated cells and the level is relatively constant when muscle mass varies. Thus, serum creatinine to cystatin C ratio can reflect the skeletal muscle mass (Kashani et al., 2017). Lower serum creatinine and cystatin C ratio is related to lower skeletal muscle mass and the higher likelihood of sarcopenia. Prior studies have reported that sarcopenia was associated with cognitive decline in elderly adults (Cabett Cipolli et al., 2019; Peng et al., 2020). However, up to now, no studies have been conducted to assess whether SI, a novel biomarker of sarcopenia, is related to cognitive function.

The aim of our study was to investigate the gender-specific correlations between SI and cognitive function using data from the China Health and Retirement Longitudinal Study (CHARLS), which was a national representative survey.

A total of 6,442 participants were included in this study and 3007 (46.7%) of them were males. The mean age was 61.5 years old (SD = 9.7) and 60.0 years old (SD = 10.0) in males and females, respectively. There were significant differences in age, BMI, education level, marital status, rural residence, smoking and alcohol drinking status, chronic diseases, hand grip strength, SI, and cognitive function scores between males and females. These characteristics were presented in Table 1.

Then, SI was treated as a categorical variable based on the quartile cut-off points in males and females, respectively (Table 2). In both males and females, we found that the mean age was lower and the mean BMI was higher in the Q4 group compared with the other groups (all P < 0.001). What’s more, the proportions of living in rural area and the low level of education were lower in the Q4 group (all P < 0.001). And the percentages of smoking were lower in the Q4 group of males (P = 0.005). Of note, the proportion of drinking was higher in the Q4 group of males (P < 0.001), while the similar phenomenon was not observed in females (P = 0.893). In particular, the hand grip strength was the greatest in the Q4 group compared with the other groups (all P < 0.001), regardless of gender. Thus, the positive relationship between SI and sarcopenia in this study was confirmed, which was consistent with the previous literatures. And the mean total cognitive function score of the Q4 group was the highest among the four groups in males, as well as in females (all P < 0.001, Figure 1).

To further investigate the relationship between SI and cognition, linear regression models were performed (Table 3). In the unadjusted model, higher SI was correlated with better cognitive function in males and females (all P < 0.001), including the three dimensions of orientation and attention, episodic memory, and visuo-construction. After adjustment for confounders, SI was significantly and positively related to better orientation and attention, episodic memory, and total cognitive function scores in males (β = 0.007, 95 CI% 0.002 to 0.012, P = 0.005; β = 0.007, 95 CI% 0.003 to 0.010, P < 0.001; β = 0.014, 95 CI% 0.007 to 0.021, P < 0.001; respectively), while higher SI was only significantly correlated with higher orientation and attention, and total cognitive function scores in females (β = 0.008, 95 CI% 0.003 to 0.013, P = 0.003; β = 0.011, 95 CI% 0.003 to 0.018, P = 0.004; respectively).

When the total cognitive function score was treated as a categorical variable according to quartiles in males and females, respectively (Q3 vs. Q1), the positive correlation between SI and cognitive function scores was observed in unadjusted logistic regression models (OR = 1.389, 95% CI 1.264 to 1.526, P < 0.001; OR = 1.480, 95% CI 1.358 to 1.614, P < 0.001; respectively; Table 4). Following adjustment for confounders, the relationship remained significant in both males and females (OR = 1.147, 95% CI 1.028 to 1.279, P = 0.014; OR = 1.219, 95% CI 1.106 to 1.344, P < 0.001; respectively; Table 4).

In this study, we found that higher SI was independently correlated with better cognitive function in both linear and logistic regression models after adjusted for confounding factors, irrespective of gender. To our knowledge, it is the first study to evaluate the relationship between SI and cognitive function.

Sarcopenia is characterized by low skeletal muscle mass and low muscle function (including muscle strength or physical performance) which occurs during normal aging (Cruz-Jentoft et al., 2010; Wiedmer et al., 2021). And recently, the diagnosis of sarcopenia is mainly based on the presence of low muscle strength and low muscle quantity or quality (Cruz-Jentoft et al., 2019). It has been reported to increase the risks of frailty, falling, cognitive impairment, functional decline, and even mortality (Cruz-Jentoft et al., 2019). A growing number of studies have proposed that SI is a credible biomarker for skeletal muscle mass (Kashani et al., 2017; Ichikawa et al., 2020; Lin et al., 2020). And this prompted us to investigate the relationship between SI and cognitive function, since SI was easy to obtain in clinic and can be widely applied.

Existing evidence has revealed that there is a clear gender difference in SI (Lin et al., 2020; Okubo et al., 2022; Tang et al., 2022). And it can be understood as the differences of serum creatinine concentration across genders since the skeletal muscle mass of males is usually greater than of females. Similarly, the diagnostic criteria for sarcopenia in males and females are also not same. Hence, we conducted the study on males and females separately.

Based on the linear and logistic regression models, evidence from our study showed that lower SI was significantly correlated with cognitive decline in males and females. To interpret this relationship, we mainly attribute it to the bidirectional effects of sarcopenia and cognition since SI is regarded as a promising biomarker for sarcopenia. The underlying mechanisms for the relationship between sarcopenia and cognitive impairment have not been fully understood. First, elevated levels of plasma pro-inflammatory cytokines are found in dementia patients, such as interleukin-6 (Licastro et al., 2000). And this could negatively influence muscle protein metabolism through direct catabolic or indirect mechanisms (Dalle et al., 2017). Besides, elderly adults with cognitive impairment are prone to stay at home and reduce food intake, which may lead to poor musculoskeletal health, like loss of muscle mass and muscle strength (Beaudart et al., 2019). Finally, reduced physical exercise, generated by sarcopenia, may decrease the level of the brain-derived neurotrophic factor which is related to learning and memory (Kang and Schuman, 1995; Erickson et al., 2011). The exact mechanisms of the relationship between SI and cognitive function remains to be elucidated by future biological studies.

A major strength of this study was a large number of participants from a nationally representative database, providing more solid conclusions. Meanwhile, this study was clinically innovative and possessed values for clinical applications. And we adjusted as much confounders as possible in our study to improve the reliability of results. However, some potential limitations in this study should to be considered. First, due to the restrictions of CHARLS database, neurocognitive functioning was not assessed by the Mini-Mental State Examination (MMSE) and the Montreal Cognitive Assessment (MoCA), which are the two most frequently used screening tools in cognition. Future studies with more formal neuropsychological assessments are required to further confirm the relationship between SI and cognitive function. Second, the exclusion of participants with incomplete data may introduce some selection bias. But ultimately, the number of included participants is large enough which can mitigate the effect of selection bias. Third, the diagnoses of participants’ chronic diseases were defined based on self-reported. And some participants may be unaware of their diseases. Confounders associated with sarcopenia, including nutrition and exercise, were not adjusted in this study since the CHARLS database didn’t provide the nutrition information and had a large proportion of missing data about exercise information. Finally, this study mainly focused on participants aged 45 years and older in China. Thus, it remained unclear whether the conclusions could be fully generalized to the populations in other countries.

In summary, we assessed the correlation of SI and cognitive function in a middle-aged and older Chinese population from the CHARLS database. And it showed that higher SI was significantly correlated with better cognitive function in both males and females. Further studies should be conducted to validate the correlation between SI and cognitive function and explore the exact mechanisms.

Publicly available datasets were analyzed in this study. This data can be found here: http://charls.pku.edu.cn/.

Ethical approval was not provided for this study on human participants because the data was from the China Health and Retirement Longitudinal Study (CHARLS), which was approved by the Biomedical Ethics Review Committee of Peking University (IRB00001052-11015). The patients/participants provided their written informed consent to participate in this study.

YZ and YY: study concept and design. YZ, ZT, SL, FZ, and CQ: acquisition, analysis, and interpretation of data. YZ: drafting the manuscript. YY and QZ: critical revision. All authors approved the final manuscript.