Postoperative delirium (POD) is a state of acute delirium caused by dysregulation of basic neuronal activity secondary to systemic disorders, and is a type of delirium with typical characteristics of consciousness fluctuations, inattention, confusion and changes in consciousness level. It is a common complication of the central nervous system after surgery, most commonly seen 1 ∼ 3 days after surgery. POD can lead to more postoperative complications, higher mortality, longer hospital stay and higher hospitalization costs. Therefore, to uncover the risk factors of POD and accurately identify the high-risk population of POD has always been one of the hot topics in perioperative research.

Cerebrospinal fluid (CSF) biomarkers include β-amyloid 42 (Aβ42), total Tau (T-tau) and phosphorylated Tau (P-tau), which are the core biomarkers of postoperative delirium (POD). Associated with the development and progression of POD, reduced CSF Aβ42 levels were shown to be an independent predictor (Strozyk et al., 2003; Lin et al., 2021), and the deposition of amyloid Aβ42 in amyloid plaques was reciprocally elevated (Blennow et al., 2015; Cunningham et al., 2019). Increased T-tau levels reflect the intensity of axonal degeneration (Hesse et al., 2001; Bitsch et al., 2002; Riemenschneider et al., 2003; Hampel et al., 2010), while increased P-tau levels are associated with the amount of neurofibrillary tangles in the brain of AD patients (Tapiola et al., 2009). Therefore, high levels of T-tau or P-tau suggest neurodegeneration (Idland et al., 2017), and their peak changes are related to the severity of POD (Parker et al., 2021). In addition, decreased Aβ/tau ratio is associated with POD (Xie et al., 2014).

Social networks refer to a system of relatively stable relationship formed by interaction between individual members of the society. They focus on the interaction and connection between people, and the consequent impact on people’s social behavior. A social network is essentially a social structure composed of many nodes, usually represented by individuals or organizations. Hence, social networks represent the social relationships ranging from casual acquaintances to tightly knit family among people and organizations.

Recently, the influence of social networks on cognitive status of middle-aged and elderly people has caught much attention. Maintaining an active social lifestyle in later life may enhance cognitive reserve and benefit cognitive function (Evans et al., 2018; Siette et al., 2020). Studies have proved that social relationship with friends or neighbors is associated with a lower risk of functional decline, and is not affected by family support (Murata et al., 2017). A rich social network was associated with a reduction in various adverse health outcomes and a lower risk of death in older adults (Oremus et al., 2019; Zahodne et al., 2021). In contrast, poor social networks may increase the risk of neurodegenerative diseases. For example, in a 2021 study, Akhter-Khan et al. (2021) reported that poor social networks (e.g., chronic loneliness) were independent predictors of AD. However, there are only a few studies on the correlation between social network and Postoperative cognitive dysfunction (POCD) and POD, and the pathological mechanism of neurodegenerative diseases caused by social network remains unclear.

Our study aims to explore whether suboptimal social networks are a risk factor for POD, and to characterize the mediating role of CSF core proteins between social networks and POD. It is of great clinical significance to explore the possibility of using social networks as an independent predictor to identify people at high risk of POD, and to validate the feasibility of using social network as an entry point to reduce the incidence of POD.

The results of this study support the association between suboptimal social networks and POD, with low social network score being an independent predictor. Social networks are founded upon interpersonal relationships, the establishment of which in turn may promote social participation and social support (Czaja et al., 2021). On the other hand, a low score for social networks may indicate relative alienation and even social isolation. In addition, animal models have proved that social isolation may cause or even aggravate memory loss (Leser and Wagner, 2015), while a rich social environment may reduce cognitive deficits in rats (Hsiao et al., 2012; Huang et al., 2015). A study conducted in 2020 demonstrated that social isolation is associated with cognitive decline in older adults (Yu et al., 2021), confirming the notion that high levels of social isolation may influence cognitive status in older adults by affecting episodic memory and mental states. Numerous studies have shown that a good social network can significantly improve the cognitive status of older adults (Fratiglioni et al., 2000, 2004; Balouch et al., 2019). People with rich social networks have more opportunities to communicate with the outside world, and are more likely to promote cognitive, physical and social participation, enhance physical and mental stimulation, and effectively reduce negative psychological states, thereby reducing the risk of cognitive dysfunction. Moreover, Preoperative cognitive status was related to POD (Choi et al., 2020), which is supported by our results. Besides, we could apply the results of this study to identify patients at a high risk for POD. Since POD, as a common perioperative complication in elderly patients, may affect the length of hospital stay and long-term prognosis, accurate preoperative identification of these patients and immediate intervention may reduce the incidence of POD, thereby shortening the length of hospital stay and improving long-term outcomes.

Based on the social network questionnaire, this study divided social networks into three subgroups. The scores of the three subcomponents reflected how close the patients were to different social relationships. Our study further examined the relationship between the scores of three subcomponents and CSF biomarkers of POD. The results confirmed that different social relationships are associated with different incidence of POD, and that the level of intimacy within the same social relationship affected CSF biomarkers.

Pathological changes of POD mainly include phosphorylation of tau protein and deposition of amyloid protein. In addition, it has been reported that the ratio between CSF biomarkers is more sensitive to the identification of amyloid deposition than individual biomarkers (e.g., Aβ42/Aβ40 vs. Aβ42) (Janelidze et al., 2016). Therefore, the ratio was also used in this study to explore the relationship between social network and POD, and it was concluded that the intimacy of the three social relationships mainly affected the incidence of POD by modulating the level of T-tau and P-tau proteins, Furthermore, the degree of intimacy with friends mainly affected the value of P-tau, Aβ42/T-tau, and Aβ42/P-tau, the degree of intimacy with neighbors mainly affected the value of P-tau, and, the degree of intimacy with relatives affected the value of Aβ42, T-tau, P-tau, Aβ42/T-tau, and Aβ42/P-tau.

In order to explore the mechanism of how social network relates to the incidence of POD, we explored the mediating effect of CSF biomarkers between social network and POD, and concluded that social network mainly affects T-tau protein, P-tau protein and the level of Aβ42/T-tau protein, Aβ42/P-tau affected the occurrence of POD. It has been found that rich social networks may reduce Aβ deposition and P-tau production (Huang et al., 2015; Donovan et al., 2016). In line with our study, previous research showed an elevated Aβ aggregation in isolated aged APP/PS1 mice, which was associated with increased secretase and decreased neprilysin expression (Hsiao et al., 2012; Huang et al., 2015). Besides, D’oleire Uquillas et al. (2018) found that after controlling for age, sex, and the Alzheimer’s disease genetic risk factor, apolipoprotein E (APOE), a higher tau pathology in the right entorhinal cortex was associated with greater loneliness. This is consistent with our research.

In this study, we established three correction models, four variables in Model 2 (age, sex, years of education, and MMSE) and nine variables in Model 3 (age, sex, MMSE, years of education, history of alcohol consumption, history of smoking, history of diabetes, history of hypertension, and history of coronary heart disease) were used as co-variables for regression analysis. Suboptimal social networks are still significantly correlated with the occurrence of POD. In addition, patients older than 65 years old were included in the regression analysis as co-variable of model 4, and the conclusion has not changed, indicating that our conclusion is robust— Suboptimal social networks is a risk factor for the occurrence of POD.

In addition to confirming the correlation between social network and the incidence of POD in the elderly population, this study is also consistent with the recent research hotspot of “age-friendly environment.” The concept of an age-friendly environment has attracted a lot of attention since its inception. How to integrate the advantages of medicine, environment, society and other disciplines and fully identify the environmental factors conducive to health and longevity is one of the cutting-edge research directions. Studies have proved that both natural and social environments have a significant impact on the physical and mental health and cognitive function of the elderly. For example, a study in 2021 confirmed that social isolation is harmful to the cognitive health of the elderly population (Li et al., 2021). Since the COVID-19 pandemic, home isolation has significantly increased the incidence of social isolation among the elderly, while unsurprisingly reducing their social support. Our result confirmed that the lack of social network is a risk factor for POD in aged population. Hence in future studies, through the preoperative score of social networks, we can predict or identify people at high risk of POD, make timely and necessary intervention, reduce the incidence of POD, so as to improve the long-term prognosis and promote the construction of healthy aging society.

There are some limitations to our study. Firstly, at the stage of data collection, we only counted three co-morbidities of patients, and the number of comorbid conditions was small. In subsequent studies, we will include as many comorbid conditions as possible to make the results more convincing. Secondly, this study is only a cross-sectional study which explores the relationship between preoperative total social network score and POD in middle-aged and elderly patients around the time of elective surgery, without following-up the incidence of delirium and cognitive function status of patients for a long time after surgery. Hence we will further explore in future studies the influence of suboptimal social networks and other social factors on longitudinal changes of POD neuropathology through multi-center CSF and neuroimaging tests with a large sample size. Finally, this study only focused on the effect of social network on POD among Chinese Han population. As different groups and regions have different cultural backgrounds, the identified effect of social network on the cognitive function and mental health status of the elderly may change. Therefore, it is necessary to sort out the dimensions of social networks, develop social network measurement tools suitable for specific cultural backgrounds, and explore the correlation between the different groups and regions of China with different cultural backgrounds.

Our study confirms that suboptimal social networks are a risk factor for POD, and that different levels of intimacy in the same social relationship can affect the concentration of CSF POD biomarkers, mainly mediated by tau protein, hence influencing the incidence of POD. In other words, it can affect cognitive function by affecting the concentration of CSF biomarkers. This suggests that suboptimal social networks, as a risk factor for POD, can predict the population at high risk. This work will provide information for future social behavior research and age-friendly environment research, thus facilitating the development of intervention strategies for POD and promoting the construction of healthy aging environment.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.

The studies involving human participants were reviewed and approved by Chinese Clinical Trial Registry (clinical registration number ChiCTR2000033439) and approved by the Ethics Committee of Qingdao Municipal Hospital. The patients/participants provided their written informed consent to participate in this study.

YB conceived the current study. FW, JW, SL, and XW performed the experiments. RD, XL, HY, and BW analyzed the data. XT and BW performed the experiments and wrote and revised the manuscript. All authors have contributed to the manuscript revising and editing critically for important intellectual content and given final approval of the version and agreed to be accountable for all aspects of the work presented here and read and approved the final manuscript.

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.