Introduction

Delirium is a common postoperative complication. postoperative delirium (POD) is transient and reversible in most cases but is associated with worse functional outcomes, increased complications, prolonged hospital stay, and mortality1–3. Numerous studies have demonstrated the detrimental effect of higher POD severity on short- and long-term outcomes4. Therefore, identifying patients at high risk for delirium and providing more attentive care to prevent POD are essential for improving postoperative clinical outcomes in the postoperative period. Multiple risk factors for POD have been identified, and Pisani et al. published an evidence-based consensus statement on the preoperative, intraoperative, and postoperative risk factors for POD5. Preoperative factors included advanced age, comorbidities, preoperative fasting and dehydration, hyponatremia or hypernatremia, and use of anticholinergic medications. Intraoperatively, the surgical site (abdomen or chest) and intraoperative bleeding were significant factors for POD. Pain was identified as a postoperative risk factor. However, there are a variety of risk factors for POD, depending on the disease and surgical technique used.

Acute cholecystitis can lead to serious complications and even death, if left untreated. For acute cholecystitis, the Tokyo Guidelines recommend cholecystectomy as soon as possible within 72 h to 1 week of onset6. However, the Tokyo Guidelines do not mention POD and the risk factors are not clear. Low BMI, a history of neuropsychiatric disease, hypo or hyperkalaemia, and prolonged operative time have been reported as significant risk factors for delirium after chronic and acute cholecystitis7. Despite the widespread use of cholecystectomy for acute cholecystitis, predictors of POD after cholecystectomy remain unclear. Therefore, this study aimed to evaluate the association between the increased incidence of POD after acute cholecystitis and the pre- and intraoperative predictors.

Materials and methods

Study design and patients

Patients who underwent cholecystectomy under general anesthesia between January 2015 and December 2020 at Kansai Medical University Medical Center were included. Patients with a pathological diagnosis of acute cholecystitis were included, and those with a diagnosis of cholelithiasis, chronic cholecystitis, and gallbladder polyps were excluded. Finally, the patients with a pathological diagnosis of acute cholecystitis were divided into POD and non-POD groups to compare their demographic characteristics and clinical outcomes.

Data collection

Patient demographics and perioperative variables were obtained from the medical records at our institution. Demographic variables included age, sex, pathology, acute cholecystitis severity6, Data on Charlson comorbidity index (CCI)8, body mass index (BMI), Subjective Global Assessment (SGA)9, Prognostic Nutritional Index (PNI = (10 × albumin (g/dL) + (0.005 × TLC: total lymphocyte count (mm3))10, American Society of Anesthesiologists (ASA) classification, number of hospital days, smoking, alcohol consumption, medical history (mental illness, diabetes mellitus (DM), hypertension (HTN), cardiovascular disease (CVD), cerebrovascular disease (CVA), pulmonary disease, renal disease), and perioperative benzodiazepine (BZD) medication use were included. Preoperative laboratory values included measures for white blood cells (WBC), C-reactive protein (CRP), aspartate transaminase (AST), alanine transaminase (ALT), gamma-GTP, serum albumin (ALB), and CRP/ALB ratio (CAR). Surgical variables included additional epidural anesthesia, time from diagnosis to surgery, surgery within 72 h of onset, operative time, blood loss, technique (open, laparoscopic, or laparotomy transition), drain use, and need for blood transfusion. Outcome measures included postoperative complications (except delirium), mortality, and length of hospital stay. POD data (calculated from the date of surgery to the date of discharge) were collected from medical and nursing records. POD was diagnosed by an experienced psychiatrist, T. F., in accordance with the delirium section of the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5).

Statistical analysis

Results are expressed as mean ± standard deviation (SD), median (interquartile range), or number (percentage), as appropriate. Chi-squared or Fisher’s exact tests were used to compare categorical data. Variables with p < 0.05 in univariate analysis were included in the multivariate analysis by backward elimination. Multiple logistic regression analysis was used to identify risk factors for POD. Receiver Operating Characteristic (ROC) curve analysis was used to determine the cutoff value of CAR for POD. p < 0.05 was considered statistically significant. All data were obtained using SPSS Statistics (version 27.0; IBM Corp., Armonk, NY, USA).

Results

Basic characteristics and intraoperative parameters

We studied 531 patients; 77 with a pathological diagnosis of acute cholecystitis were included (Fig. 1). The patients were divided into POD (n = 18) and non-POD (n = 59) groups. The demographic data and perioperative variables of the study population are presented in Table 1. The total incidence of POD was 23.3% (18/77). The mean age of the patients was 75.6 ± 11.772 (range: 44–93) years for the POD group and 69.2 ± 13.292 (range: 27–93) years for the non-POD group. In the POD group, 18 patients (100%) had a history of alcohol consumption, whereas 45 patients (76.3) had a history of alcohol consumption in the non-POD group, a significant difference between the two groups (P = 0.022). For SGA, the measures were 2.56 ± 0.86 in the POD group and 1.86 ± 0.8 in the non-POD group, a significant difference between the two groups (P = 0.002). The POD group was discharged 16.56 ± 16.41 days postoperatively and the non-POD group was discharged after 13.88 ± 11.43 days, with no significant difference (p = 0.438).

Fig. 1
figure 1

Flowchart of the inclusion and exclusion process for patient enrollment in the study. A total of 531 patients were enrolled, and 77 patients who underwent cholecystectomy for acute cholecystitis were divided into the POD (n = 18) and non-POD (n = 59) groups.

Full size image
Table 1 Demographic and clinical parameters of the POD and non-POD groups.
Full size table

Laboratory measurements

The preoperative CRP and ALB levels and CAR for patients in the POD group were 23.36 ± 10.76 mg/L, 2.72 ± 0.61 g/L, and 8.97 ± 5.82, respectively, while those in the non-POD group had CRP of 14.37 ± 10.82 mg/L, ALB of 3.38 ± 0.76 g/L, and CAR of 4.71 ± 3.88. There were statistically significant differences in CRP and ALB levels and CAR between the two groups (all p < 0.001) (Table 2).

Table 2 The laboratory test of the POD and non-POD groups.
Full size table

Risk factors for POD

Variables with p < 0.05 in univariate analysis (alcohol consumption, history of psychiatric disease, SGA score, and CAR) were included in the multivariate analysis by backward elimination. Finally, eight factors (age, sex, alcohol consumption, CCI, history of psychiatric disease, SGA score, CAR, and operative time) were included in the multivariate logistic regression model (forward selection (likelihood ratio), p < 0.001, Nagelkerke’s R2 = 0.382). Results showed that CAR (β = 0.15; odds ratio = 1.161; 95% confidence interval (CI), 1.006–1.341; P = 0.041), SGA score (β = 0.953; odds ratio = 2.546; 95% CI, 1.05–6.174; P = 0.039), and history of psychiatric disease (β = 1.568; odds ratio = 4.796; 95% CI, 1.276–18.026; P = 0.02) were the three independent risk factors for POD (Table 3).

Table 3 Risk factors for POD.
Full size table

Predictive value of CAR for POD

The predictive value of preoperative CAR for POD was evaluated by ROC analysis. As shown in Fig. 2, the area under the curve of CAR for POD was 0.731, with a cutoff value of 3.69, sensitivity of 94.1%, and specificity of 43.9% (95% CI, 0.608–0.853; P = 0.004). The positive likelihood ratios (+ LR) and negative likelihood ratios (− LR) were 1.676 and 0.134, respectively, which of the Area Under the Curve (AUC) value of CAR (AUC: 0.731) were superior to those of CRP (AUC: 0.709) and ALB (AUC: 0.26). Based on a cutoff value of 3.69, patients were classified into a high CAR group (CAR ≥ 3.69) and a low CAR group (CAR < 3.69) (Table 4).

Fig. 2
figure 2

(A) Comparison of CARs between POD and non-POD groups. POD, postoperative delirium; CAR, C-reactive protein to albumin ratio. P < 0.05 was considered statistically significant. (B) The predictive value of CAR for POD by ROC curve analysis. The AUC of CAR for POD was 0.731, with a cut-off value of 3.69, a sensitivity of 94.1%, and a specificity of 43.9% (95% CI: 0.608–0.853, P = 0.004).

Full size image
Table 4 Comparison of the AUC for the three inflammation-based prognostic scores.
Full size table

Discussion

In terms of basic characteristics and intraoperative parameters, SGA score, alcohol consumption, and history of psychiatric disorders differed significantly between POD and non-POD groups. In laboratory measurements, there were significant differences between POD and non-POD groups in CRP and ALB levels and CAR. There were no significant differences in other background factors. In the Tokyo Guideline, mortality, complication rates, bile duct injury rates, and open conversion rates were lower for early (72 h to 1 week) cholecystectomy for acute cholecystitis compared to standby surgery6. However, there was no significant difference in the time elapsed from the onset of acute cholecystitis to surgery between the two groups. Furthermore, patients were classified according to whether surgery was performed within 72 h, and there was no difference in the incidence of POD when surgery was performed within 72 h of onset.

The results of current study in the multivariate logistic regression model showed that the CAR and SGA score were higher in patients with POD than in those without. Furthermore, patients with psychiatric disorders were at higher risk of developing POD. Malnutrition and a history of neuropsychiatric disorders have been reported as significant risk factors for POD after chronic and acute cholecystitis7.

Neuroinflammatory changes in the central nervous system have been proposed as an explanatory mechanism for the background factors of POD11. CRP level is one of the most common biomarkers for systemic inflammation. Some studies have shown that the CRP level is an independent risk factor for POD after hip surgery12, vascular surgery13, and laparoscopic surgery for colon cancer14. Another study reported high serum CRP levels preoperatively and on postoperative day 2 as potential predictors of POD in elderly patients after major noncardiac surgery.15 ALB level is frequently used to assess the nutritional status of patients undergoing surgery. Previous studies have shown that hypoalbuminemia is significantly associated with an increased risk of POD16, 17. Several studies have reported that severe preoperative hypoalbuminemia is a predictor of POD and worse outcomes in patients undergoing noncardiac surgery18. CRP and ALB levels can also predict morbidity, mortality, and poor outcomes, such as longer hospitalization and intensive care unit (ICU) stays, respectively; elevated CRP levels are associated with malignancy, sepsis, and inflammatory disease; and decreased ALB levels are associated with pre-existing medical conditions, liver failure, renal failure, and malnutrition due to pre-existing hepatic, and renal conditions and malnutrition19, 20. Preoperative nutritional status correlates with postoperative complications, including POD, strongly suggesting the importance of improving nutritional status as much as possible before surgery.

In contrast to changes in CRP and ALB levels alone, which are nonspecific as each is associated with multiple pathologies, CAR correlates with prognostic potential by more accurately reflecting the severity of nutrient deprivation and inflammation19,20,21. Previous studies have demonstrated the predictive ability of the CAR for morbidity, mortality, and other outcomes in various patients, including critically ill19, pre-transplant or cirrhotic22,23,24, postoperative20, 25, 26, and oncological patients.27, 28 The CAR has also been reported as a preoperative predictive indicator of POD in total knee29, 30 and hip30, 31 arthroplasty surgeries in elderly patients. In the present study, a preoperative CAR of 3.69 or higher was a risk factor for POD in acute cholecystitis with surgery.

POD occurs on postoperative days 2–53, 16 and prolongs hospital stay by 2–3 days and intensive care unit stay by 2 days32, 33 POD is associated with 7–10% surgery-related mortality, but has been reported to be 1% in patients without delirium34. In the POD group, all patients experienced disease onset within 24 h. In our study, there was no difference in surgery-related mortality, complications, or length of hospital stay. There were no cases of rehospitalization in either group. However, because early detection of POD is expected to reduce medical resources and increase patient well-being, a new preoperative evaluation of CAR may be beneficial as an assessment measure of the risk factor for the development of POD. Since CRP and ALB are common items in preoperative blood tests for patients with acute cholecystitis, we believe that there is a significant advantage to adding CAR as a preoperative evaluation item, as there is no direct burden on the patient or additional medical resources. For patients predicted to develop POD based on the calculated CAR values, multidisciplinary POD countermeasures from preoperative to immediate postoperative period may contribute to the reduction of factors that promote POD.

POD is caused by multiple factors and presents with various clinical syndromes and pathophysiological changes. Our understanding of delirium captures only a subset of these symptoms. Although many previous studies show an association between POD and inflammation indices, few have been disease-limited studies. By limiting the disease, specific background factors included in previous studies can be excluded. And since there is no literature on POD in only acute cholecystitis. To our knowledge, this is the first study to report preoperative CAR as a predictor of POD in patients undergoing cholecystectomy. A limitation of this study is that it was a retrospective, single-center study that needs to be validated in a prospective, multicenter study. Furthermore, the predictors were measured only preoperatively, ignoring their relationship with the longitudinal improvement or worsening of delirium for each patient. Some residual confounding factors (for example, sample selection bias and preoperative comorbidities) cannot be completely excluded. In particular, CRP may contribute to the improvement or worsening of delirium, as it may increase or decrease between the preoperative and postoperative periods. We improved the clinical pathway starting in 2022 to standardize the schedule of blood draws and items to observe the course of POD and biomarker variability. And by adding new blood collection items, we plan to conduct a more in-depth study for the next report. It would be interesting to investigate the dynamic changes in the CAR after surgery in future studies to investigate POD outcomes and CAR dynamics. On the other hand, the application of CAR to other acute diseases and hepatobiliary-pancreatic surgery with highly invasive may contribute by adding new depth to existing literature. Because this study was retrospective, the definition and severity of delirium may not have been accurately identified. The patients with low-activity delirium were included in the non-delirium group. It is unclear whether CAR in combination with other variables (such as IL-6 and TNF-α) can predict the risk of POD, and further studies are needed to validate the same. As both CRP and ALB are synthesized in the liver, an improved assessment of the exact relationship between liver dysfunction and serum CAR levels in acute cholecystitis is also a topic for future study.

Conclusions

In conclusion, preoperative CAR, preexisting psychiatric disorders, and SGA score may be promising predictors of POD after cholecystectomy for acute cholecystitis. In particular, the new preoperative evaluation of CAR may be beneficial as an assessment measure of the risk factor for the development of POD.