Preoperative CD11c^hi atypical memory B cells are associated with postoperative delirium

The work complies with the Transparency in the Reporting of Artificial Intelligence (TITAN) Guidelines 2025[1].

Advanced age is an independent risk factor for postoperative delirium (POD), but POD does not occur in all elderly patients, so it is important to characterize disease-related features in high-risk populations^[2,3]. Immunosenescence, a hallmark of aging, has recently been found to lead to severe adaptive immune dysregulation and autoimmune disorders, and both have been shown to play a role in the pathogenesis of various types of neurocognitive disorders, such as rheumatic disease-associated neuropsychiatric manifestations, autoantibody-associated encephalitis, Alzheimer’s disease, and neurological symptoms following vaccination^[4-7]. However, the association between aging-related adaptive immune dysregulation and perioperative neurocognitive disorders remains limited. A key immunosenescence signature is the accumulation of CD11c^hi atypical memory B cells (AMBCs), defined by high CD11c expression with concurrent CD21/CD27 deficiency[4]. This distinct B cell subset demonstrates remarkable efficiency in autoantibody production, coupled with potent cytotoxic activity and robust proinflammatory capacity[8]. These pathogenic characteristics may increase susceptibility to severe perioperative complications under surgical stress in high-risk populations. The present study aims to investigate the association between preoperative CD11c^hi AMBC levels and POD in elderly surgical patients.

HIGHLIGHTS

• Elevated preoperative levels of CD11c^hi atypical memory B cells (AMBCs) were independently associated with postoperative delirium (POD) incidence.

• Preoperative levels of CD11c^hi AMBCs were significantly correlated with the severity of POD.

• Preoperative CD11c^hi AMBCs serve as novel blood-based biomarkers for early POD risk warning.

Following approval by the Institutional Ethics Committee of Hunan Provincial People’s Hospital, China (#2024334), clinical trial registration (ChiCTR2400090345), and obtaining written informed consent, we conducted a prospective observational study between November 2024 and March 2025. A total of 90 patients were finally enrolled (Supplementary Figure 1, available at: http://links.lww.com/JS9/E394), with comprehensive perioperative data collected including patient demographics, intraoperative parameters, and postoperative outcomes (Supplementary Table 1, available at: http://links.lww.com/JS9/E393). Cognitive function was assessed preoperatively using the Mini-Mental State Examination (MMSE)[9]. Delirium screening was performed using the Confusion Assessment Method at two time points: (1) prior to leaving the operating room during recovery, and (2) daily from 8:00 AM to 12:00 noon on postoperative days 1–7[9]. Delirium severity was quantified in all patients using the Memorial Delirium Assessment Scale (MDAS)[10]. Peripheral venous blood samples (5 mL) were obtained preoperatively from the upper extremities and processed within 24 hours for quantification of circulating CD11c^hi AMBCs (Supplementary Figure 2, available at: http://links.lww.com/JS9/E394).

All included participants received standardized anesthesia protocols without preoperative medication. Protocol 1: Combined spinal-epidural anesthesia performed by experienced anesthesiologists at L3–4 interspace using a 16-gauge Tuohy needle for epidural puncture, followed by intrathecal administration of 15 mg 0.5% ropivacaine via 25-gauge Whitacre needle and epidural catheter placement, supplemented with propofol infusion (1–2 mg kg⁻¹ h⁻¹) for basic sedation. Protocol 2: General anesthesia induction with propofol (1 mg kg⁻¹), sufentanil (0.25–0.45 μg kg⁻¹), and vecuronium (0.1 mg kg⁻¹), maintained with remifentanil (0.1–0.2 μg kg⁻¹ min⁻¹), propofol (1–2 mg kg⁻¹ h⁻¹), and 1–2% sevoflurane in 50% oxygen, maintaining heart rate and blood pressure within 20% of baseline values and BIS 40–60. Postoperative analgesia included hydromorphone (0.2 mg kg⁻¹ in 100 mL saline) via continuous infusion (2 mL h⁻¹) for 48 hours with 0.5 mL bolus doses (10-min lockout interval). Rescue analgesia (flurbiprofen axetil 50 mg IV) was administered for visual analogue scale (VAS) scores ≥3.

Patients were grouped based on whether POD (POD vs non-POD) occurred. Statistical analysis was performed with SPSS, version 27.0 (Statistical Package for Social Sciences). The normal distribution of the variables was examined using the Kolmogorov–Smirnov test. Continuous data were presented as mean (SD) and compared using the unpaired t-test if distributed normally. Data that were not normally distributed were reported as median (IQR) and analyzed using the Mann–Whitney test. Categorical variables were reported as numbers (%) and compared using χ2 or Fisher exact test, as appropriate. Receiver operating characteristic (ROC) curve was plotted and the diagnostic efficacy was assessed by computing the numeric value of the area under the curve (AUC). Binary logistic regression models were employed to assess the association between CD11c^hi AMBC levels and POD, with POD as the dependent variable and CD11c^hi AMBC levels as the independent variable. The model was further adjusted for the interaction between biomarkers, age, sex, preoperative MMSE scores, and education status. Association estimates were expressed as odds ratios (OR) or β-coefficients with corresponding 95% confidence intervals (CI). Two-sided P <0.05 was considered to be statistically significant.

The overall incidence of POD was 22.2% (20/90) among all 90 patients (Supplementary Table 1, available at: http://links.lww.com/JS9/E393). The participants who developed POD had higher preoperative levels of CD11c^hi AMBCs (1.5 (1.31, 1.82) versus 0.99 (0.68, 1.33), P <0.001) than participants who did not develop the POD (Supplementary Table 1, available at: http://links.lww.com/JS9/E393 and Fig. 1A and B). The AUC for CD11c^hi AMBCs was 0.778, indicating an accurate diagnostic efficacy (Fig. 1C). Results of the unadjusted analyses and multivariable models are presented in Supplementary Table 2 (available at: http://links.lww.com/JS9/E393). After adjustment for preoperative MMSE, the levels of CD11c^hi AMBCs (OR: 3.20, 95% CI: 1.28–8.75, P = 0.016) were still associated with POD. In the crude model, an increased level of CD11c^hi AMBCs was associated with an increase in POD severity. This association persisted after adjustment for preoperative MMSE (β coefficient: 0.97, 95% CI: 0.23–1.71, P = 0.011; Supplementary Table 2, available at: http://links.lww.com/JS9/E393).

Figure 1.

(A) Representative images and quantification of CD11c^hi AMBCs in the POD and non-POD participants. (B) Different preoperative levels of CD11c^hi AMBCs between the participants with POD and those without POD. Participants who developed POD (n = 20) had higher preoperative levels of CD11c^hi AMBCs than the participants who did not develop POD (n = 70). The Mann-Whitney test was used to analyze the data. The data were presented as median with interquartile range. (C) ROC curve of preoperative levels of CD11c^hi AMBCs in predicting POD. Abbreviations: AMBCs, atypical memory B cells; AUC, area under the curve; POD, postoperative delirium; ROC, receiver operating characteristic.

This study demonstrates that elevated preoperative CD11c^hi AMBC levels are independently associated with POD incidence and severity after adjusting for confounders. Meanwhile, preoperative levels of CD11c^hi AMBCs showed good diagnostic efficacy for POD. These findings suggest that CD11c^hi AMBCs may serve as novel blood-based biomarkers for early POD risk warning and potentially contribute to POD pathogenesis. However, these conclusions are limited by the restricted sample size and require validation through larger multi-center cohorts. The precise mechanisms through which CD11c^hi AMBCs mediate neurocognitive impairment warrant further investigation.

Ethical approval

For the clinical study protocol, it was set in compliance with Helsinki Declaration and was provided by the Institutional Ethics Committee of Hunan Provincial People’s Hospital, China (#2024334), Changsha, China.

Consent

Informed consent was obtained from all the subjects involved in the study.

Sources of funding

This work was supported by grants from the Hunan Provincial Department of Education (24C0028 to S.T.), Guangxi Science and Technology Base and Talent Special Project (AD25069060 to Y.X.), Guangxi Key Research and Development Program (AB24010066 to Y.X.), and National Natural Science Foundation of China (82460239 to Y.X.).

Author contributions

S.T., L.W., and Y.X.: conceptualization; S.T., W.C., Y.Z., and X.Y.: data curation; S.T., W.C., Y.Z., and X.Y.: formal analysis; S.T. and Y.X.: funding acquisition; S.T., W.C., Y.Z., X.Y., methodology; S.T., L.W., and Y.X.: project administration; S.T., J.L., Y.Y., C.T., and M.R.: software; S.T., X.Y., J.L., and Y.Y.: visualization; B.P., and G.K.: validation; S.T., L.W., and Y.X.: supervision; S.T. and W.C.: writing – original draft; S.T., Y.Z., X.Y., L.W., and Y.X.: writing – review & editing.

Conflicts of interest disclosure

The authors declare no conflicts of interest.

Guarantor

Lai Wei and Yubo Xie.

Research registration unique identifying number (UIN)

The trial was registered in the Chinese Clinical Trial Registry (ChiCTR2400090345; Principal Investigator: S.T.; Date of Registration: 27 September 2024).

Provenance and peer review

Not commissioned, externally peer-reviewed.

Data availability statement

The datasets used and analyzed during this study are available from the corresponding author upon reasonable request.