Due to the progressive aging of the global population, it is anticipated that the number of hip fractures worldwide will increase to an estimated 6.1 million cases by 2050, with an annual escalation of 1 to 3% (1). Postoperative delirium (POD) is a frequent and severe complication in geriatric patients who have undergone hip fracture surgery, with an incidence rate ranging from 4 to 53% (2–10). POD is an acute cognitive dysfunction characterized by changes in consciousness or attention that cannot be attributed to pre-existing cognitive impairment (2, 8, 9). Hallucinations and delusions, restlessness, disturbed speech, decreased attention, disordered thinking, agitated behavior, and disturbances in sleep–wake cycles are the most prevalent symptoms (5, 7). POD usually occurs within the first 5 days after surgery, predominantly within the first 48 h (7). Numerous studies have demonstrated that POD has a significant impact on surgical outcomes, cognitive recovery, and healthcare costs (2). Furthermore, POD is associated with prolonged hospital stays, increased mortality rates at 6 months and 1 year, and the development of permanent cognitive impairment (5, 11).

Multiple prospective studies have demonstrated that hypoalbuminemia can contribute to the occurrence of POD. In one such study, Venkatakrishnaiah et al. (10) prospectively evaluated 110 geriatric hip fracture patients and found that hypoalbuminemia was a strong predictor of POD. Similarly, Ishihara et al. (12) conducted a multicenter prospective study and showed that serum albumin concentration ≤ 37 g/L was an independent risk factor for POD in patients undergoing initial hepatectomy. Zhang et al. (13) investigated preoperative hypoalbuminemia in a clinical trial involving geriatric ICU patients after surgery and reported an increased risk of POD and poor prognosis. Finally, a multicenter prospective study conducted by Matsuki et al. (14) revealed a correlation between low serum albumin levels and the incidence of POD in urological surgery patients.

While several studies (2, 5–10, 15) have confirmed the relationship between preoperative serum albumin levels and POD, the general and oversimplified conclusions have limited the specificity of clinical practices and hindered the promotion of preoperative albumin management. Currently, focusing on modifiable risk factors for intervention represents the most effective approach to prevent adverse outcomes of POD (6, 16). This study aims to employ a retrospective cohort design to meticulously classify preoperative serum albumin levels and investigate the strength of the association between different levels of preoperative serum albumin and POD. The objective is to provide clinicians with precise and quantitative predictive associations and determine the threshold level of preoperative albumin that benefits patients. This research will aid in enhancing the feasibility of clinical interventions and offer new evidence for the routine assessment and management of preoperative albumin levels in hip fracture patients.

This study seeks to address several critical questions through its retrospective cohort design:

From January 2013 to November 2023, a total of 2031 electronic medical records were collected. After applying strict inclusion and exclusion criteria, a retrospective cohort study was conducted on 1,440 patients. Among them, 325 patients (22.6%) had serum albumin levels below 35 g/L, and 275 patients (19.1%) experienced POD. The study population had an average hospital stay of 12 days. The selection process is illustrated in Figure 1.

Table 1 presents the basic characteristics of the patients, categorized according to the severity of preoperative hypoalbuminemia. The distribution of patients in the normal albumin level group, mild, moderate, and severe hypoalbuminemia groups was as follows: 1115 patients (77.4%), 216 patients (15.0%), 70 patients (4.9%), and 39 patients (2.7%), respectively. The average age of all included patients was 75 years, with males accounting for 39.7%. Among the patients, 6.0% had dementia, 24.5% had diabetes mellitus. The Jonckheere-Terpstra test demonstrated a statistically significant trend in various variables as serum albumin levels increased. These variables include age, Frailty index, dementia, preoperative delirium, ASA classification, time to surgery, intraoperative blood loss, transfusion, lymphocyte count, and blood glucose level. Moreover, please consult Table 1 for the revised results obtained through the application of the Bonferroni method. Additionally, compared to the normal albumin level group, the hypoalbuminemia group had a higher incidence of POD (Figures 2A, p < 0.001), and the incidence of POD increased with the severity of hypoalbuminemia (Figure 2B; trend p < 0.001).

Multiple logistic regression analysis, adjusted for covariates (see Supplementary Table 3 for details), was conducted, and the results are presented in Table 2. When the preoperative albumin level was treated as a continuous variable, the association between preoperative albumin level and POD was demonstrated with an odds ratio (OR) of 1.11 (95% CI: 1.08–1.15, p < 0.001). In other words, for every 1 g/L decrease in albumin level, the risk of POD increased by 11%. When categorized into two groups, the hypoalbuminemia group had an OR of 2.99 (95% CI: 2.14–4.18, p < 0.001) compared to the normal albumin level group. According to the severity of hypoalbuminemia, the adjusted ORs were 2.71 (95% CI: 1.84–3.99) for mild hypoalbuminemia, 3.44 (95% CI: 1.88–6.28) for moderate hypoalbuminemia, and 3.97 (95% CI: 1.78–8.86) for severe hypoalbuminemia, with a trend p < 0.001.

Furthermore, 1:1 propensity score matching was performed between the normal albumin level group and the hypoalbuminemia group to balance all covariates (SMD < 0.1). The baseline characteristics of patients before and after matching are displayed in Supplementary Table 4. Logistic regression analysis of the matched cohort showed that as the albumin level decreased, the risk of POD gradually increased, with a significant association between the two, as detailed in Table 2.

Figure 3A presents the restricted cubic spline plot depicting the relationship between preoperative albumin level and POD. The model was adjusted for the clinical severity of albumin and all included covariates. The results indicated a linear dose–response relationship, showing that as the preoperative albumin level decreased, the risk of POD increased (P for Nonlinear = 0.59). Additionally, when the preoperative albumin level fell below the critical value of 38.00 g/L, the risk of POD significantly increased, with an odds ratio of 1. Furthermore, Figure 3B illustrated the association between albumin level and the probability and incidence of POD, highlighting that lower preoperative albumin levels were linked to an increased incidence of POD.

In subgroup analysis, Figure 4 depicted the interaction between preoperative albumin level (normal albumin level vs. hypoalbuminemia) and POD, along with other covariates. The results revealed significant interactions between diabetes and hypoalbuminemia (interaction p < 0.001). Specifically, patients without diabetes were more prone to develop POD in the presence of preoperative hypoalbuminemia. Therefore, clinicians should be particularly vigilant about POD occurrence when managing patients without diabetes.

POD is a prevalent and significant issue among patients with traumatic hip fractures. While identifying and correcting preoperative risk factors has been a hot topic of concern, current mainstream research (2, 7, 18) has primarily focused on identifying independent risk factors that are difficult to modify or intervene upon, such as advanced age, male gender, dementia, diabetes, ASA classification, and inflammatory indicators such as neutrophil count, lymphocyte count, and C-reactive protein. Recent studies have explored various inflammatory indicators, with Kim et al. (6) suggesting that the C-reactive protein/albumin ratio can predict the occurrence of POD. However, limited research has been conducted on easily modifiable preoperative independent risk factors that are clinically manageable and promotable, such as preoperative serum albumin and blood glucose levels. This research gap hinders clinical practitioners from accurately predicting and intervening in POD caused by preoperative hypoalbuminemia in patients with hip fractures. Therefore, it is crucial to comprehensively understand the varying levels of risk associated with different albumin levels. Doing so will provide valuable insights that can aid clinical doctors in identifying high-risk POD patients early and formulating targeted intervention measures (10).

In our study, through stratified handling of serum albumin levels and multifactorial logistic regression analysis, we obtained precise estimates of the strength of the association between different albumin levels and POD. We found a linear dose–response relationship between serum albumin levels and POD in hip fracture patients: the lower the albumin level, the higher the incidence of POD. Moreover, when the preoperative albumin level was below 38.00 g/L, the risk of POD significantly increased. This finding is consistent with Cabrerizo et al. (19), who also indicated an elevated risk of postoperative complications in geriatric hip fracture patients with albumin levels below 38 g/L. Additionally, subgroup analysis demonstrated a noteworthy interaction between hypoalbuminemia and diabetes in individuals with hip fractures. Nevertheless, cautious interpretation of the interaction effect on diabetes is warranted, considering factors such as sample size. Moreover, the association between preoperative diabetes and preoperative hypoalbuminemia requires further investigation and validation in future studies.

Although the precise mechanism underlying POD remains unknown, various theoretical hypotheses have been proposed in mainstream research to explain its pathogenesis (20–22). These include alterations in central neurotransmitters, decreased brain metabolic levels, inadequate cerebral perfusion, and neuroinflammation. The neuroinflammation hypothesis proposes that external stressors such as surgery and trauma can trigger a systemic immune inflammatory response (6, 23). During a systemic immune inflammatory response, excessive expression of pro-inflammatory cytokines activates vascular endothelial cells and perivascular cells at the blood–brain barrier (6, 24). This disrupts or increases the permeability of the blood–brain barrier, enhancing the transport of peripheral inflammatory cytokines to the brain (9), subsequently activating microglial cells to generate an inflammatory response, interfering with synaptic connections and transmission (7, 25), and even leading to neuronal ischemia and apoptosis (6, 26). In summary, inflammation plays a crucial role in the occurrence of POD.

Albumin is commonly deficient in the older (27, 28). Considering serum albumin levels solely as a nutritional indicator is narrow-minded. Serum albumin not only reflects nutritional status but also closely relates to antioxidant capacity, free radical scavenging, and other functions (15, 29). Albumin is widely distributed in blood, interstitial fluid, and cells, accounting for three-quarters of plasma antioxidant capacity (30, 31). Studies have shown that the antioxidant and free radical scavenging effects of albumin depend on its strong ligand-binding properties, enabling it to bind with free fatty acids, Cu2+, Fe3+, etc., preventing their peroxidation reactions and the formation of reactive oxygen species (30, 32). Additionally, the cysteine at position 34 in albumin contains free thiol, which can scavenge hydroxyl radicals (33).

Taylor et al. (34) conducted a study demonstrating the association between POD and disruption of the blood–brain barrier. They utilized the cerebrospinal fluid/plasma albumin ratio (CPAR) as a reliable measure of blood–brain barrier permeability. Their findings revealed that the severity of delirium was positively correlated with an increase in CPAR, suggesting a link between delirium and blood–brain barrier breakdown. Similarly, Devinney et al. (35) employed a similar methodology and observed that patients who were experiencing delirium exhibited a significantly greater increase in CPAR from preoperative to postoperative 24 h, compared to those who did not experience delirium. Based on these results, they concluded that the heightened permeability of the blood–brain barrier after surgery is independently associated with the occurrence of POD. These changes in blood–brain barrier permeability not only facilitate the transportation of inflammatory factors but also lead to the escape of albumin from the serum into the brain, where it acts as a major extracellular scavenger and antioxidant, perhaps serving as a beneficial mechanism to prevent excessive inflammation and avoid damage to the body (30).

However, in the central nervous system, low serum albumin levels appear to be insufficient in fully exerting their antioxidant effects and capturing free radicals (10, 36, 37). This defense mechanism is hindered in the central nervous system, which can lead to cognitive impairment (10, 38). Additionally, normal serum albumin levels can provide the cell and immune system with essential amino acids and nitrogen-containing substances necessary for body construction, thereby optimizing the composition ratio between albumin, acute phase proteins, and other inflammatory factors to enhance the body’s immune response capabilities (39–43).

Although there is controversy surrounding the benefits of exogenous supplementation of serum albumin, based on existing knowledge, we must consider the impact of hypoalbuminemia on inflammatory activity and the risks it portends (44). This aligns with previous research showing that nutritional supplementation can reduce the occurrence of acute mental disorders during acute trauma and postoperatively and improve clinical outcomes for patients (5, 45–47). It has been widely acknowledged that early identification and intervention in modifiable risk factors are currently the most effective strategy for preventing and managing POD’s complexity (10, 48). Furthermore, assessing preoperative serum albumin levels assists clinicians in identifying high-risk POD patients at an earlier stage. Thus, the objective of our study is to encourage clinicians to emphasize the assessment and management of preoperative serum albumin levels.

In patients undergoing surgical treatment for hip fractures, we observed a linear dose–response relationship between preoperative serum albumin levels and the incidence of POD. Our findings suggest that maintaining a preoperative serum albumin level above 38 g/L may lead to more favorable outcomes. Consequently, we recommend that clinicians adopt standardized protocols for managing preoperative serum albumin levels, highlighting the significant impact of hypoalbuminemia on inflammatory activity and predictive risk. In order to further investigate the potential benefits, it will be crucial to conduct randomized controlled trials assessing the outcome effects of tightly controlling preoperative serum albumin levels in patients with hip fractures.

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

The studies involving humans were approved by Ethics Committee of Dandong Central Hospital (Approval no. DDZX-20231101). The studies were conducted in accordance with the local legislation and institutional requirements. The ethics committee/institutional review board waived the requirement of written informed consent for participation from the participants or the participants’ legal guardians/next of kin because This is a retrospective cohort study that did not involve the use of patients’ names, addresses, or any other personally identifiable information.

WW: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Writing – original draft, Writing – review & editing. WY: Writing – review & editing. WT: Writing – review & editing. YL: Writing – original draft. QL: Writing – review & editing. WD: Writing – review & editing.