Multivariable models of postoperative delirium in cardiac surgery patients: proteomic and demographic contributions

Gonçalves MCB, Khera T, Otu HH, Narayanan S, Dillon ST, Shanker A, Gu X, Jung Y, Ngo LH, Marcantonio ER, Libermann TA, Subramaniam B
doi: https://doi.org/10.1101/2023.05.30.23289741

ABSTRACT

Background: Delirium following cardiac surgery is common, morbid, and costly, but may be prevented with risk stratification and targeted intervention. Preoperative protein signatures may identify patients at increased risk for worse postoperative outcomes, including delirium. In this study, we aimed to identify plasma protein biomarkers and develop a predictive model for postoperative delirium in older patients undergoing cardiac surgery, while also uncovering possible pathophysiological mechanisms. Methods: SOMAscan analysis of 1,305 proteins in the plasma from 57 older adults undergoing cardiac surgery requiring cardiopulmonary bypass was conducted to define delirium-specific protein signatures at baseline (PREOP) and postoperative day 2 (POD2). Selected proteins were validated in 115 patients using the ELLA multiplex immunoassay platform. Proteins were combined with clinical and demographic variables to build multivariable models that estimate the risk of postoperative delirium and bring light to the underlying pathophysiology. Results: A total of 666 proteins from SOMAscan analysis were found altered between PREOP and POD2 (Benjamini-Hochberg (BH)-p<0.01). Using these results and findings from other studies, 12 biomarker candidates (Tukey’s fold change (|tFC|)>1.4) were selected for ELLA multiplex validation. Eight proteins were significantly altered at PREOP, and seven proteins at POD2 (p<0.05), in patients who developed postoperative delirium compared to non-delirious patients. Statistical analyses of model fit resulted in selecting a combination of age, sex, and different 3-protein biomarker panels highly correlated with delirium at PREOP (angiopoietin-2, ANGPT2; C-C motif chemokine 5, CCL5; and metalloproteinase inhibitor 1, TIMP1; AUC=0.829) and POD2 (lipocalin-2, LCN2; neurofilament light chain, NFL; and CCL5; AUC=0.845). The delirium-associated proteins identified as biomarker candidates are involved with inflammation, glial dysfunction, vascularization, and hemostasis, highlighting the multifactorial pathophysiology of delirium. Conclusion: Our study proposes two models of postoperative delirium that include a combination of older age, female sex, and altered levels of proteins both preoperatively and postoperatively. Our results support the identification of patients at higher risk of developing postoperative delirium after cardiac surgery and provide insights on the underlying pathophysiology. ClinicalTrials.gov (NCT02546765).

CLINICAL PERSPECTIVES

  • Comprehensive proteomics screening of patients undergoing cardiac surgery and its association with the development of postoperative delirium.

  • We propose two different multivariable models: the first model inputs information on risk factors for developing postoperative delirium after cardiac surgery, which could be used clinically to preoperatively predict those patients at higher risk; and the second model informs on postoperative proteomic signatures that bring light to delirium pathophysiology and underlying mechanisms for further investigation.

INTRODUCTION

Delirium is a multifaceted condition characterized by an acute and fluctuating decline in attention and cognitive function. Highly prevalent after surgery, it is a perioperative neurocognitive disorder (NCD)1 that affects up to 50% of older adults undergoing cardiac surgery2. Postoperative delirium has been associated with short and long-term cognitive impairment that can last from a few weeks to as long as 7.5 years after surgery1. Additional delirium-associated adverse outcomes include postoperative complications, functional decline, and increased mortality1–4. From the institutional perspective, delirium prolongs ICU and hospital stay, increases the need for institutional discharge, and results in an inflated cumulative cost of care ($44,291 higher costs per patient per year) when compared to non-delirious surgical patients5.

The development of postoperative delirium is multifactorial. Predisposing factors include age and previous medical history6, while precipitating factors include surgery-related stressors (e.g. type and duration of surgical procedures, anesthetic medications, intraoperative hypotension, and hypoxia), as well as postoperative pain and intensive care management. Combined, those factors contribute to pathophysiological mechanisms such as a pro-inflammatory environment, which has become one of the hypotheses behind the development of NCDs7–10. Surgery outcome and recovery can be negatively impacted by the sustained inflammation resulting from the surgical trauma, which elicits stress responses and the prolonged activation of both innate and adaptive immune systems11, 12. Previous studies supporting the inflammation hypothesis for postoperative delirium have shown, for example, greater baseline and postoperative levels of C-reactive protein (CRP), and fluctuating levels of IL-6 during the course of delirium after surgery13–16.

Clinical risk prediction scales and models have been proposed as tools to identify and stratify the overall risk of developing neurocognitive complications postoperatively17–23. However, cardiac surgical patients remain under-represented among the targeted populations24. Despite the advances and efforts from the past decade21, 25, identifying risk factors for postoperative delirium in patients undergoing cardiac surgery remains challenging26.

The development and constant upgrading of high throughput “omics” technologies and systems biology emerge as a potent tool for biomarker research focused on developing multivariable models to predict complex clinical conditions such as postoperative delirium27. Therefore, in this study we conducted a comprehensive proteomic profiling and screening of biomarker candidates to predict postoperative delirium in surgical cardiac patients. By combining relevant demographic variables and peripheral protein levels from participants of the DexAcet trial28, we present two multicomponent models for postoperative delirium in older patients undergoing cardiac surgery.

METHODS

Study design and population

The DexAcet trial, a prospective, randomized, placebo-controlled, triple blinded, factorial design trial, was carried out as a pilot study at a single tertiary care center, from September 2015 through April 2018, with long-term neurocognitive follow-ups completed in April 201929. This study was approved by the Committee on Clinical Investigations Institutional Review Board (CCI-IRB Protocol # 2019 P00075) at Beth Israel Deaconess Medical Center (BIDMC; Boston, MA) and registered at ClinicalTrials.gov (NCT02546765). The details of the protocol and results were previously published28, 30. Briefly, the DexAcet study enrolled 120 patients aged 60 years and above undergoing cardiac surgery requiring cardiopulmonary bypass (CPB), including Coronary Artery Bypass Grafting (CABG) with or without valve replacement surgery. Exclusion criteria included emergent procedure, isolated aortic surgery, preoperative ejection fraction <30%, preexisting cognitive impairment, Alzheimer’s disease (AD), Parkinson’s disease (PD), on treatment for cognitive decline, history of recent seizures, serum creatinine levels above 2 mg/dL, liver dysfunction, recent history of alcohol misuse, English language limitation, or hypersensitivity to any study medication. Patients were randomly assigned into four arms (1:1:1:1) to receive either intravenous (I.V.) acetaminophen along with propofol or dexmedetomidine, or a placebo along with propofol or dexmedetomidine. Acetaminophen (1g I.V.) or placebo was administered upon immediate arrival at the ICU and continued for eight doses, given six-hourly for 48 hours. A sensitivity analysis showed that the results here reported were not affected by the study procedures and drugs administered.

The flow chart for the current study is summarized in Figure 1.

Figure 1.
Figure 1. Study flowchart

A total of 115 patients of the DexAcet clinical trial were included in this study to identify protein signatures for postoperative delirium in patients undergoing cardiac surgery requiring cardiopulmonary bypass. BH = Benjamini-Hochberg; tFC = Fold-change of Tukey’s biweight robust averages; NCDs = perioperative neurocognitive disorders.

Medical record review

The medical record was reviewed daily by trained research associates and was the primary source for patient demographics, details of the intraoperative and postoperative course, and discharge status.

Delirium and cognitive assessments

Baseline and discharge cognitive assessments were performed in all subjects using the Montreal cognitive assessment (MoCA) as previously described31. Delirium was assessed on each postoperative day by trained research associates who administered a brief cognitive battery to assess attention, orientation, and memory, and the Delirium Symptom Interview32 to probe for delirium-related symptoms such as unclear thinking and perceptual disturbances. Based on these data, the presence of delirium was determined using the Confusion Assessment Method diagnostic algorithm33. For further details, see Subramaniam et al., 201928.

Biological assessments

Blood samples were collected before surgery (PREOP) and at postoperative day 2 (POD2). Once collected, samples were fractionated into components (red cells, buffy coat, and plasma). The aliquoted plasma and buffy coat samples were securely stored in locked –80 °C freezers and later transferred to the lab for further proteomic analyses, ensuring cold chain maintenance. The 115 participants from the DexAcet trial who completed the two-timepoints blood collection were included in the current proteomic study. None of the results from these studies were shared with participants or their family members. All patients approached for participation in the study also agreed to participate in blood collection/storage as noted in the informed consent.

SOMAscan assay

Considering that propofol is the most common sedative used in our local Cardiac ICU, and no significant difference was observed in delirium outcome according to sedative choice (dexmedetomidine vs. propofol) in the DexAcet trial28, PREOP and POD2 samples from the subgroup who received I.V. propofol for sedation (57 participants) were selected for the proteomic screening. Plasma samples (50 DL) from the 57 patients, five pooled plasma controls, and one buffer control were analyzed using the SOMAscan Assay Kit for human plasma 1.3k, measuring expression of 1,305 proteins. The assay was performed at the BIDMC Genomics, Proteomics, Bioinformatics, and Systems Biology Center according to the manufacturer’s standard protocol (SomaLogic; Boulder, CO). Several hybridization controls further controlled sample-to-sample variability. Each protein concentration was transformed into a corresponding SOMAmer concentration, then quantified using a custom DNA microarray (Agilent; Santa Clara, CA) read-out, which reports the data as relative fluorescence units (RFU). Data quality control, normalization, and calibration were done according to the manufacturer’s protocol as previously described34.

Selection of biomarker candidates

A total of twelve biomarker candidates were selected according to the following criteria:

  1. Significant differential protein expression observed from the SOMAscan proteomic scanning, as a result of:

    1. Surgery, with a defined p-value cut-off and Tukey’s fold change (tFC) cut-off of POD2 vs. PREOP, Benjamini-Hochberg (BH)-p<0.01, |tFC|>1.4; and

    2. Delirium, at either PREOP or POD2, with a defined p-value cut-off of p<0.05 for Delirium vs. No delirium; and/or

  2. Literature review of proteins for their association with inflammation, cognitive decline, AD, delirium, and other NCDs; and

  3. Current availability of antibody pairs to be analyzed by a 4-plex-based ELLA immunoassay.

ELLA immunoassay

Plasma levels of the twelve proteins selected as biomarker candidates from the SOMAscan screening were measured in the 115 patients from the DexAcet trial28 using the fully automated immunoassay platform ELLA (ProteinSimple/Biotechne; San Jose, CA) and custom 4-plex immunoassays optimized for ELLA, following the manufacturer’s protocols. Inter-assay coefficients of variation (CVs) of triplicate measures were generally <5%. If a CV was >10%, the assay was repeated. Assay personnel were blinded to case and control status.

Statistical Analyses

SOMAscan proteome analysis and identification of biomarker candidates

Paired t-tests were applied to the SOMAscan proteome data to identify proteins significantly differentially expressed between POD2 and PREOP with a BH-corrected paired t-test p-value less than 0.01. FC was calculated as the ratio of the one-step Tukey’s bi-weight average of signal values in the two groups (POD2 and PREOP) to obtain a robust estimate of the average expression level of a protein that is not affected by outliers. Sample and feature clustering was done using principal component analysis (PCA) and hierarchical clustering with average linkage using Pearson’s correlation as the distance metric. Support Vector Machines (SVM) based on PCA results was used for sample classification using linear, polynomial, and Gaussian kernels. The same set of analyses was performed to define the dysregulated proteins in delirium cases vs. no delirium cases at PREOP and POD2. However, due to the subtler differences for delirium and imbalanced sample sets, the threshold for significance was raised to a nominal p-value lower than 0.05.

Statistical analysis of ELLA immunoassay data

Normality assumptions of protein concentrations and clinical outcomes at the different timepoints were assessed via Shapiro-Wilk test prior to any statistical data comparison. Protein levels (pg/mL) were overall not normally distributed among the total population (n=115; delirium = 21; no delirium = 94) and results are shown as median and interquartile ranges. Comparisons of mean level of protein between delirium and no delirium within each time point (PREOP, POD2) were performed using unpaired t-test for near normally distributed data, and Mann-Whitney non-parametric test for non-normally distributed data. The difference in surgical effect (mean level between PREOP and POD2) between delirium and no delirium group was assessed by testing the interaction term between delirium status (yes/no) and the protein concentration in a general linear model.

Development of multivariable models

To arrive at the selected models that incorporated an optimal set of baseline demographics and clinical variables, and the protein biomarkers, we took into consideration the following factors. First, with only 21 delirium cases, and with the rule of thumb of having at minimum 5 positive outcomes per predictor in the multivariable model, we limited the dimension of the model to at most 5 predictors. We considered age, sex, and MoCA as our baseline demographics and clinical predictors of interest. We then used all subset model selection method by adding into this base model all combinations of the proteins (e.g. one at a time, two at time, three at a time) but limited up to three proteins and assessed all these models that have the base variables of age, sex, and MoCA with 0 up to 3 proteins. We used two criteria to rank these models: area under the receiver operating characteristic curve (AUC) which was estimated by the concordance statistic (c-statistic), and the standard model selection criterion of the Akaike Information Criterion (AIC). The larger the AUC the better the discrimination between Delirium and No Delirium groups. The smaller the AIC the better the goodness-of-fit of the model. We used both criteria to select the optimal set of models. We favored first selecting the model with the largest AUC values, then among them, see which models yielded the smallest AIC. These criteria avoid the pitfalls of multiple testing which exist in model selection criteria such as forward, backward, stepwise. Using AUC, and AIC allowed us to evaluate and weigh clinical interpretation in selecting or recommending the chosen model(s). This model selection process was used for both PREOP and POD2 data separately. All analyses were conducted in SAS, Version 9.4, Cary NC.

RESULTS

Study population

A total of 115 participants from the DexAcet trial28 were included in this study (Table 1). Participants had median age of 69 years (IQR, 63-76), were mostly male (97/115, 84.3%), white race (106/115 [92.2%]), and had baseline MoCA scores of 23 [IQR, 21-26]). Among these 115 participants, 21 (18.3%) experienced postoperative delirium. Patients who developed delirium were older than those who did not develop delirium (median age of 75 [IQR, 69-77] vs. 69 [IQR, 63-73] years, respectively; p=0.047) and females were more likely to develop delirium than males (38.9% vs. 14.4%, p=0.03). The most common comorbidities among all patients were hypertension (100/115, 87%), dyslipidemia (97/115, 84.3%), and diabetes (49/115, 43%), although no direct association between these conditions and delirium outcome was observed. Isolated coronary artery bypass grafting (CABG) was the most applied surgical procedure in our study population (71/115, 61.7%), followed by CABG + Valve Replacement (33/115, 28.7%).

View this table:
Table 1. Study population characteristics

SOMAscan-proteomic profiling of cardiac patients

The SOMAscan proteomic scanning was performed on plasma samples of 57 participants, of which 12 (21.05%) experienced postoperative delirium (Supplemental Table S1). From the 1,305 proteins scanned, we found significantly altered levels of 666 analytes after cardiac surgery at POD2 compared to PREOP (BH-p<0.01) (Supplemental Table S2). PCA-SVM and hierarchical clustering analyses demonstrated clear discrimination between PREOP and POD2 timepoints (Figure 2A and 2B). When comparing the proteomic profile of delirious and non-delirious patients, we measured differential expression of 115 and 85 proteins at PREOP and POD2, respectively (nominal-p<0.05) (Supplemental table S3). The separation of two groups along the principal component 1 on the PCA-SVM analyses (Figure 2C and 2D) suggests that delirium was a major contributor to the differential protein levels at both timepoints.

Figure 2.
Figure 2. Differential and qualitative analyses of SOMAscan-based global proteome screening of 57 cardiac surgery patients requiring cardiopulmonary bypass

(A). The heatmap for unsupervised hierarchical clustering of 666 differentially expressed proteins among patients shows differences between POD2 and PREOP (BH-p<0.01). (B) Principal component analysis of all 57 patients based on the 666 differentially expressed proteins projected in the 2D plot shows a clear segregation of features. Component 1 splits the samples by surgery states (POD2 vs. PREOP). (C) Principal component analysis of all 57 patients based on the 115 differentially expressed proteins at PREOP projected in the 2D plot shows a clear segregation of features. Component 1 splits the samples by delirium effect (DEL vs NON-DEL). (D) Principal component analysis of all 57 patients based on the 85 differentially expressed proteins at POD2 projected in the 2D plot shows a clear segregation of features. Component 1 splits the samples by delirium effect (DEL vs NON-DEL).

Selection of postoperative delirium biomarker candidates

From the 666 differentially expressed analytes (BH-p<0.01), 225 showed |tFC|>1.4 (Supplemental Table S2), 12 final targets were selected as biomarker candidates following the criteria described in methods for further validation by orthogonal immunoassays of the surgery effect and delirium outcome: platelet-derived growth factor subunit B (PDGFB), tumor necrosis factor receptor superfamily member 1A (TNFRSF1A/TNF-sR1A/TNFR1/CD120a), interleukin-6 (IL6), C-C motif chemokine 5 (CCL5/RANTES), insulin-like growth factor-binding protein 2 (IGFBP2), metalloproteinase inhibitor 1 (TIMP1), angiopoietin-2 (ANGPT2), chitinase-3-like protein 1 (CHI3L1/YKL-40), hepcidin (HAMP), thrombospondin-1 (THBS1), neurofilament light chain (NfL), and lipocalin-2 (LCN2). These twelve proteins are overall moderately correlated at both PREOP and POD2 (Supplemental Table S4a and S4b, respectively), and are involved in several pathways relevant for delirium, such as inflammation, vascular function, neuronal injury, and extracellular matrix degradation.

Immunovalidation of protein levels and the association with delirium

All 12 ELLA-measured proteins in 115 patients were highly significant (p<0.01) for the surgery effect (9 proteins increased, 3 proteins decreased at POD2 vs PREOP), correlating well with the SOMAscan data (data not shown). Additionally, 8 out of the 12 ELLA-measured proteins were significantly associated with delirium at PREOP: increased IGFBP2, ANGPT2, IL6, TIMP1, CHI3L1, NFL, TNFRSF1A; decreased PDGFB (delirium vs no delirium; p<0.05; Figure 3). Similarly, 7 of the 12 proteins were associated with delirium at POD2: increased IGFBP2, LCN2, IL6, TIMP1, NFL, TNFRSF1A; decreased CCL5 (delirium vs no delirium; p<0.05; POD2; Figure 3). No statistically significant differences were observed at PREOP or at POD2 for HAMP (p=0.842; p=0.115) and THBS1 (p=0.126; p=0.066) protein levels between the delirium and no delirium groups. Further stratification according to the four study-arms from DexAcet trial demonstrated overall minimal effect of drug treatments on protein levels, justifying the pooled analysis presented here (data not shown). Analyses of changes in protein concentrations between PREOP and POD2 stratified by delirium status (Figure 3) showed greater increases in LCN2 (p=0.01) and TNFRSF1A (p=0.02) levels in the delirium vs. no delirium group, and less increases in HAMP (p=0.03) in the delirium vs. no delirium group. Overall, the magnitude of surgical changes was similar for all other biomarker candidates in the delirium and no delirium groups (DiD; ΔΔ Perioperative).

Figure 3.
Figure 3. Quantitative validation of biomarker candidates by ELLA 4-multiplex immunoassay according to delirium and surgery status, and the difference in changes in protein levels between delirious and non-delirious patients during the perioperative period

Protein levels (pg/mL) of the twelve selected biomarker candidates were analyzed according to delirium experience among all the 115 patients undergoing cardiac surgery. Scatter plots represent median and interquartile range of plasma protein concentration; p-values correspond to the difference of protein levels between delirium and no delirium outcomes at each timepoint separately (PREOP and POD2). Difference-in-differences (ΔΔ) regression analysis represents the relative protein changes during the perioperative period and are represented by the shading plots; p-values correspond to the difference between protein levels variation during surgery according to delirium status. ANGPT2 = angiopoietin-2; CCL5 = C-C motif chemokine 5; CHI3L1 = chitinase-3-like protein 1; HAMP = hepcidin; IGFBP2 = insulin-like growth factor-binding protein 2; IL6 = interleukin-6; LCN2 = lipocalin 2; NFL = neurofilament light chain; PDGFB = platelet-derived growth factor subunit B; POD2 = postoperative day 2 timepoint; PREOP = preoperative baseline timepoint; THBS1 = thrombospondin-1; TIMP1 = metalloproteinase inhibitor 1; TNFRSF1A = tumor necrosis factor receptor superfamily member 1A. (*p < 0.05; **p < 0.01).

Multivariable models for postoperative delirium after cardiac surgery

Multivariable analyses were performed for two different purposes by including demographic characteristics and up to five proteins’ levels. First, we aimed to identify a predictive risk model by analyzing the data collected preoperatively (PREOP). The best model based on AIC and AUC to preoperatively predict patients at higher risk of developing delirium was achieved when combining age, sex, and altered preoperative levels of TIMP1, ANGPT2, and CCL5 (PREOP; AIC=101.68, AUC=0.829; Table 2a). Second, we aimed to identify a possible proteomic signature linked to the delirium outcome by analyzing the data collected postoperatively (POD2). For that, the best model to explore the postoperative delirium pathophysiology included age, sex, and altered levels of LCN2, CCL5, and NFL at day 2 after cardiac surgery (POD2; AIC=92.04, AUC=0.845; Table 2b). Characteristics of the tested multivariable models linked to delirium risk and pathophysiology at PREOP and POD2 are summarized in Table 3a and 3b, respectively.

View this table:
Table 2a. Preoperative Delirium prediction models with demographic and clinical features, and a combination of proteins
View this table:
Table 2b. Postoperative day 2 delirium prediction model fits with demographic and clinical features, and a combination of proteins
View this table:
Table 3a. Summary of results from the best PREOP model
View this table:
Table 3b. Summary of results from the best POD2 model

DISCUSSION

In this study, we performed SOMAscan proteomics to identify potential plasma biomarkers for postoperative delirium in patients undergoing cardiac surgery. Applying orthogonal immunoassays to further validate proteomics-based delirium biomarker candidates, we derived 5-component logistic regression models to 1) predict postoperative delirium preoperatively (PREOP); and 2) identify major components of the pathophysiology of delirium once the condition is established after surgery (POD2). The two best models include both age and sex, and a combination of three peripheral blood biomarkers. Key predictive proteins included in our models mapped to the inflammatory, vascular function, and neuronal injury pathways, suggesting an important role of these processes in the onset and development of delirium. Moreover, pending further validation, the preoperative model might be useful to risk stratify older adults undergoing cardiac surgery for delirium.

From the demographic perspective, we found that age is associated with increased risk of delirium, which is consistent with the previous delirium literature35. On the other hand, we found that, even though males represent the larger portion of patients undergoing cardiac surgery, females are more likely to experience postoperative delirium, which differs from the delirium literature. For instance, Wang et al. (2021) reported a statistical predictive model where males younger than 60 are at higher risk of developing postoperative delirium after cardiac surgery36. Male sex is also reported as a risk factor in other non-cardiac surgical risk models37, 38.

Here it is important to highlight that CABG can be more challenging in females compared to males when taking into consideration surgical characteristics such as the number of vessels bypassed and the pump time. Additionally, females present smaller arteries, which are often technically more difficult to bypass. Thus, we might hypothesize that, while females get CABG less frequently, the surgery itself is considered more difficult, which might lead to more delirium. Nevertheless, our discrepant results compared to the current literature encourage further investigation in larger and more diverse populations, with more robust internal control for variants linked to technical, clinical, biological, and sociodemographic differences among the targeted population.

Though to our knowledge no protein signature has been successfully established for postoperative delirium after cardiac surgery, previous studies have looked for reliable biomarkers in patients undergoing non-cardiac surgery. For example, tau protein and NfL have been proposed as potential candidates as delirium biomarkers due to their well-known association with cognitive decline and the shared characteristics between NCDs, neurodegenerative diseases, and neuronal injury39. Using the SOMAscan technology, the Successful Aging after Elective Surgery (SAGES) study identified chitinase-3-like-protein-1 (CHI3L1/YKL-40) as a delirium-associated protein after elective non-cardiac surgeries, primarily orthopedic, and validated this protein across the entire SAGES cohort using the ELLA immunoassay platform14, 40. We also found CHI3L1/YKL-40 to be dysregulated at PREOP in cardiac surgery patients who developed delirium, though it was not selected for inclusion in our final multivariable predictive model.

In addition, our analyses show that about half of all analytes investigated (666 out of 1,305) are significantly altered by surgery, as observed from comparing protein expression at POD2 and PREOP. This scenario corroborates recent findings on clinical trials investigating both cardiac and non-cardiac patients regarding the impact of surgical trauma on specific protein targets linked to neurocognitive impairment, brain injury and delirium41–44 plus the recent SOMAscan analysis of the SAGES study40.

The stress-immune hypothesis for delirium and postoperative cognitive decline has also been supported by evidence showing increased levels of cortisol, CRP, TNF-α, IL2, IL6, and IL8 in delirious surgical patients13, 42, 44–49. The systemic inflammation resulting from the surgical trauma is also believed to induce neuroinflammation and affect brain function and cognition, especially in older and frail patients10. In this context, neuroinflammation encompasses an imbalance of systemic pro- and anti-inflammatory processes, evidenced by postoperative complement system activation, induction of the coagulation cascade, increased blood-brain barrier (BBB) permeability followed by the infiltration of immune cells, and activation of glial cells10. We identified, besides age and sex, eight proteins at PREOP and seven proteins at POD2 as potential biomarkers for postoperative delirium which may contribute to delirium risk or development.

In terms of mechanisms, in addition to inflammation, our proteomic analysis points to impaired neovascularization (especially angiogenesis), coagulation, and hemostasis, accompanied by glial dysfunction in delirious patients. From the clinical applicability perspective, the predictive models described in this paper may be helpful in two ways. First, if further validated, the PREOP predictive model may be used to risk stratify older adults undergoing cardiac surgery for delirium prevention interventions. Second, the dysregulated proteins, particularly in the POD2 model, might be useful for identifying therapeutic targets to develop interventions to either prevent or treat delirium. Prior to clinical use, both models need independent validation and further refinement in larger populations, considering both additional clinical and protein risk factors.

LIMITATIONS

It is important to note a few limitations regarding this study. The DexAcet trial was a multifactorial randomized controlled trial focusing on the protective effect of acetaminophen on postoperative delirium and thus not a pure observational study. Patients who received acetaminophen may not have developed delirium after surgery. Even though our sensitivity analysis did not find major differences on protein levels when comparing the four arms of the original study, the low incidence of delirium in the overall study must be acknowledged. For example, the SOMAscan proteomics screened participants that received propofol (n=57), however the orthogonal validation with automated immunoassays also included participants that received dexmedetomidine. Although our sensitivity analyses did not detect any significant influence of the anesthetic of choice on the proteomic level, this is an important aspect to be cautiously followed up in further studies with larger sample size.

It is common practice to observe BH-corrected proteins in data analysis. However, in the context of delirium effect, the observed effect size is comparatively subtle and our dataset is imbalanced. Therefore, BH-adjusted significance is not apparent. To overcome this limitation, we have employed a nominal p-value cut-off approach to generate a comprehensive list of proteins. From this list, candidate biomarkers are selected for further validation using the ELLA technique. Despite the small sample size and data imbalance limitations, this approach ensures that our final signature of differentially expressed proteins is not solely based on statistical analysis of SOMA data, but also includes experimental validation, thereby minimizing the inclusion of false positives.

Noticeably, our PREOP model presents calculated confidence intervals that cross 1, indicating that individual biomarkers are not significant, which might result from our small sample size. This effect is reduced in the POD2 model, suggesting a stronger significance of this model on the studied population. Hence, the sample size was relatively small and restricted to one site, and further validation of these models in a larger and diverse population will be important to confirm clinical utility. We also acknowledge that we did not consider some of previously validated risk factors, such as history of depression, stroke, and abnormal albumin levels. Although we have conducted an analysis of correlation for each risk factor and demographic data from our cohort, we did not find significant results – which might be explained by the small sample size.

Additionally, this work was conducted using SOMAscan version 3.2, which sampled 1,305 plasma proteins. The most recent SOMAscan version 4.1 measures 7,000 proteins, which could potentially identify other surgery and delirium-associated markers missed by our current approach. Also, for technical limitations, our selection criteria of the twelve biomarker candidates included only proteins that were available to be validated by the ELLA platform. Thus, the models and pathophysiological insights here proposed could benefit from more broad and detailed pathway analyses and expanded immunoassay validation, and the data presented in Table 2 might be of interest for further exploration.

CONCLUSION

Surgeries are invasive procedures that can lead to acute stress and systemic inflammatory responses associated with several adverse postoperative outcomes, including delirium. Our study shows that a combination of higher age, female sex and altered levels of 3 proteins both preoperatively and postoperatively can identify patients at higher risk of developing delirium after cardiac surgery. These predictive models, if validated in future studies, set the stage for implementing effective targeted delirium risk reduction strategies for patients undergoing cardiac surgery.

Data Availability

All data produced in the present study are available upon reasonable request to the authors.

Sources of Funding

The study was funded in part by Mallinckrodt Pharmaceuticals to conduct the SOMAscan analysis.

Disclosures

Dr Subramaniam, and Dr Marcantonio reported receiving grant support from Mallinckrodt Pharmaceuticals during the conduct of the DexAcet trial. Drs. Subramaniam and Marcantonio are currently funded by NIH-R01AG065554. Dr. Marcantonio and Dr Libermann are funded by NIH R01-AG051658 and P01-AG031720. Dr. Marcantonio is also funded by NIH K24-AG035075. Mallinckrodt Pharmaceuticals had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication. The authors had sole authority for the data, analysis, write-up, and submission.

Acknowledgments

The authors acknowledge all clinical and research staff that contributed to the study procedures.

Footnotes

  • ↵† Co-senior authors.

REFERENCES

  1. 1.↵
    Evered L, Silbert B, Knopman DS, Scott DA, DeKosky ST, Rasmussen LS, et al. Recommendations for the Nomenclature of Cognitive Change Associated with Anaesthesia and Surgery-2018. Anesthesiology 2018;129:872–879.
    OpenUrlCrossRefPubMed
  2. 2.↵
    Marcantonio ER. Delirium in Hospitalized Older Adults. The New England journal of medicine 2017;377:1456–1466.
    OpenUrlCrossRefPubMed
  3. 3.
    Inouye SK, Westendorp RGJ, Saczynski JS. Delirium in elderly people. Lancet (London, England) 2014;383:911–922.
    OpenUrl
  4. 4.↵
    Rudolph JL, Marcantonio ER. Postoperative Delirium: Acute Change with Long-Term Implications. Anesthesia and analgesia 2011;112:1202.
    OpenUrlCrossRefPubMed
  5. 5.↵
    Gou RY, Hshieh TT, Marcantonio ER, Cooper Z, Jones RN, Travison TG, et al. One-Year Medicare Costs Associated With Delirium in Older Patients Undergoing Major Elective Surgery. JAMA surgery 2021;156:430–442.
    OpenUrl
  6. 6.↵
    Bramley P, McArthur K, Blayney A, McCullagh I. Risk factors for postoperative delirium: An umbrella review of systematic reviews. Int J Surg 2021;93:106063.
  7. 7.↵
    Hshieh TT, Vasunilashorn S, D’Aquila M, Arnold SE, Dickerson B, Fong TG, et al. The Role of Inflammation after Surgery for Elders (RISE) study: Study design, procedures, and cohort profile. Alzheimer’s & dementia (Amsterdam, Netherlands) 2019;11:752–762.
    OpenUrl
  8. 8.
    Liu X, Yu Y, Zhu S. Inflammatory markers in postoperative delirium (POD) and cognitive dysfunction (POCD): A meta-analysis of observational studies. PLOS ONE 2018;13:e0195659.
  9. 9.
    Safavynia SA, Goldstein PA. The role of neuroinflammation in postoperative cognitive dysfunction: Moving from hypothesis to treatment. Frontiers in Psychiatry 2019;9:752.
  10. 10.↵
    Yang T, Velagapudi R, Terrando N. Neuroinflammation after surgery: from mechanisms to therapeutic targets. Nature Immunology 2020 21:11 2020;21:1319–1326.
    OpenUrl
  11. 11.↵
    Cerejeira J, Lagarto L, Mukaetova-Ladinska EB. The immunology of delirium. Neuroimmunomodulation 2014;21:72–78.
    OpenUrl
  12. 12.↵
    Gaudillière B, Fragiadakis GK, Bruggner R V., Nicolau M, Finck R, Tingle M, et al. Clinical recovery from surgery correlates with single-cell immune signatures. Science Translational Medicine 2014;6.
  13. 13.↵
    Dillon ST, Vasunilashorn SM, Ngo L, Otu HH, Inouye SK, Jones RN, et al. Higher C-Reactive Protein Levels Predict Postoperative Delirium in Older Patients Undergoing Major Elective Surgery: A Longitudinal Nested Case-Control Study. Biological psychiatry 2017;81:145–153.
    OpenUrl
  14. 14.↵
    Vasunilashorn SM, Dillon ST, Chan NY, Fong TG, Joseph M, Tripp B, et al. Proteome-Wide Analysis Using SOMAscan Identifies and Validates Chitinase-3-Like Protein 1 as a Risk and Disease Marker of Delirium Among Older Adults Undergoing Major Elective Surgery. J Gerontol A Biol Sci Med Sci 2022;77:484– 493.
    OpenUrl
  15. 15.
    Vasunilashorn SM, Dillon ST, Inouye SK, Ngo LH, Fong TG, Jones RN, et al. High C-reactive Protein Predicts Delirium Incidence, Duration, and Feature Severity after Major Non-cardiac Surgery. Journal of the American Geriatrics Society 2017;65:e109.
  16. 16.↵
    Vasunilashorn SM, Ngo L, Inouye SK, Libermann TA, Jones RN, Alsop DC, et al. Cytokines and Postoperative Delirium in Older Patients Undergoing Major Elective Surgery. J Gerontol A Biol Sci Med Sci 2015;70:1289–1295.
    OpenUrlCrossRefPubMed
  17. 17.↵
    Inouye SK, Zhang Y, Jones RN, Kiely DK, Yang F, Marcantonio ER. Risk factors for delirium at discharge: development and validation of a predictive model. Archives of internal medicine 2007;167:1406–1413.
    OpenUrlCrossRefPubMedWeb of Science
  18. 18.
    Inouye SK, Viscoli CM, Horwitz RI, Hurst LD, Tinetti ME. A predictive model for delirium in hospitalized elderly medical patients based on admission characteristics. Annals of internal medicine 1993;119:474–481.
    OpenUrlCrossRefPubMedWeb of Science
  19. 19.
    Kim EM, Li G, Kim M. Development of a risk score to predict postoperative delirium in hip fracture patients. Anesthesia and analgesia 2020;130:79.
  20. 20.
    Marcantonio ER, Goldman L, Mangione CM, Ludwig LE, Muraca B, Haslauer CM, et al. A clinical prediction rule for delirium after elective noncardiac surgery. JAMA 1994;271:134–139.
    OpenUrlCrossRefPubMedWeb of Science
  21. 21.↵
    Rudolph JL, Jones RN, Levkoff SE, Rockett C, Inouye SK, Sellke FW, et al. Derivation and validation of a preoperative prediction rule for delirium after cardiac surgery. Circulation 2009;119:229–236.
    OpenUrlAbstract/FREE Full Text
  22. 22.
    Shanker A, Upadhyay P, Rangasamy V, Muralidhar K, Subramaniam B. Impact of Frailty in Cardiac Surgical Patients—Assessment, Burden, and Recommendations. Annals of Cardiac Anaesthesia 2021;24:133.
  23. 23.↵
    Wassenaar A, van den Boogaard M, van Achterberg T, Slooter AJC, Kuiper MA, Hoogendoorn ME, et al. Multinational development and validation of an early prediction model for delirium in ICU patients. Intensive care medicine 2015;41:1048–1056.
    OpenUrlPubMed
  24. 24.↵
    Lindroth H, Bratzke L, Purvis S, Brown R, Coburn M, Mrkobrada M, et al. Systematic review of prediction models for delirium in the older adult inpatient. BMJ Open 2018;8.
  25. 25.↵
    Menzenbach J, Guttenthaler V, Kirfel A, Ricchiuto A, Neumann C, Adler L, et al. Estimating patients’ risk for postoperative delirium from preoperative routine data - Trial design of the PRe-Operative prediction of postoperative DElirium by appropriate SCreening (PROPDESC) study - A monocentre prospective observational trial. Contemporary Clinical Trials Communications 2020;17:100501.
  26. 26.↵
    Ramspek CL, Jager KJ, Dekker FW, Zoccali C, Diepen M van. External validation of prognostic models: what, why, how, when and where? Clinical Kidney Journal 2021;14:49.
  27. 27.↵
    Glaab E, Rauschenberger A, Banzi R, Gerardi C, Garcia P, Demotes J. Biomarker discovery studies for patient stratification using machine learning analysis of omics data: a scoping review. BMJ open 2021;11:e053674.
  28. 28.↵
    Subramaniam B, Shankar P, Shaefi S, Mueller A, O’Gara B, Banner-Goodspeed V, et al. Effect of Intravenous Acetaminophen vs Placebo Combined With Propofol or Dexmedetomidine on Postoperative Delirium Among Older Patients Following Cardiac Surgery: The DEXACET Randomized Clinical Trial. JAMA 2019;321:686.
  29. 29.↵
    Khera T, Helfand J, Kelly L, Mueller A, Shankar P, Marcantonio ER, et al. Twelve-Month Cognitive and Functional Outcomes Following Cardiac Surgery: The DEXACET Trial of Intravenous Acetaminophen Versus Placebo. Front Pharmacol 2022;13:803903.
  30. 30.↵
    Shankar P, Mueller A, Packiasabapathy S, Gasangwa D, Patxot M, O’Gara B, et al. Dexmedetomidine and intravenous acetaminophen for the prevention of postoperative delirium following cardiac surgery (DEXACET trial): protocol for a prospective randomized controlled trial. Trials 2018;19.
  31. 31.↵
    Nasreddine ZS, Phillips NA, Bédirian V, Charbonneau S, Whitehead V, Collin I, et al. The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment. Journal of the American Geriatrics Society 2005;53:695–699.
    OpenUrlCrossRefPubMedWeb of Science
  32. 32.↵
    Albert MS, Levkoff SE, Reilly C, Liptzin B, Pilgrim D, Cleary PD, et al. The delirium symptom interview: an interview for the detection of delirium symptoms in hospitalized patients. Journal of geriatric psychiatry and neurology 1992;5:14–21.
    OpenUrlCrossRefPubMed
  33. 33.↵
    Inouye SK, Van Dyck CH, Alessi CA, Balkin S, Siegal AP, Horwitz RI. Clarifying confusion: the confusion assessment method. A new method for detection of delirium. Annals of internal medicine 1990;113:941–948.
    OpenUrlCrossRefPubMedWeb of Science
  34. 34.↵
    Gold L, Walker JJ, Wilcox SK, Williams S. Advances in human proteomics at high scale with the SOMAscan proteomics platform. New biotechnology 2012;29:543– 549.
    OpenUrl
  35. 35.↵
    Ormseth CH, LaHue SC, Oldham MA, Josephson SA, Whitaker E, Douglas VC. Predisposing and Precipitating Factors Associated With Delirium: A Systematic Review. JAMA Network Open 2023;6:e2249950.
  36. 36.↵
    Wang H, Guo X, Zhu X, Li Y, Jia Y, Zhang Z, et al. Gender Differences and Postoperative Delirium in Adult Patients Undergoing Cardiac Valve Surgery. Frontiers in Cardiovascular Medicine 2021;0:1737.
    OpenUrl
  37. 37.↵
    Oh ES, Sieber FE, Leoutsakos J-M, Inouye SK, Lee HB. Sex Differences in Hip Fracture Surgery: Preoperative Risk Factors for Delirium and Postoperative Outcomes. J Am Geriatr Soc 2016;64:1616–1621.
    OpenUrlPubMed
  38. 38.↵
    Wittmann M, Kirfel A, Jossen D, Mayr A, Menzenbach J. The Impact of Perioperative and Predisposing Risk Factors on the Development of Postoperative Delirium and a Possible Gender Difference. Geriatrics 2022;7:65.
  39. 39.↵
    Fong TG, Vasunilashorn SM, Ngo L, Libermann TA, Dillon ST, Schmitt EM, et al. Association of Plasma Neurofilament Light with Postoperative Delirium. Annals of neurology 2020;88:984.
  40. 40.↵
    Fong TG, Chan NY, Dillon ST, Zhou W, Tripp B, Ngo LH, et al. Identification of Plasma Proteome Signatures Associated with Surgery Using SOMAscan. Annals of surgery 2021;273:732.
  41. 41.↵
    McKay TB, Rhee J, Colon K, Adelsberger K, Turco I, Mueller A, et al. Preliminary Study of Serum Biomarkers Associated With Delirium After Major Cardiac Surgery. Journal of Cardiothoracic and Vascular Anesthesia 2022;36:118–124.
    OpenUrl
  42. 42.↵
    Vasunilashorn SM, Ngo LH, Chan NY, Zhou W, Dillon ST, Otu HH, et al. Development of a Dynamic Multi-Protein Signature of Postoperative Delirium. The journals of gerontology Series A, Biological sciences and medical sciences 2019;74:261–268.
    OpenUrl
  43. 43.
    Lindblom RPF, Shen Q, Axén S, Landegren U, Kamali-Moghaddam M, Thelin S. Protein Profiling in Serum and Cerebrospinal Fluid Following Complex Surgery on the Thoracic Aorta Identifies Biological Markers of Neurologic Injury. Journal of Cardiovascular Translational Research 2018;11:503.
  44. 44.↵
    Rhee J, Kuznetsov A, McKay T, Lyons M, Houstis N, Mekkonen J, et al. Serum Proteomics of Older Patients Undergoing Major Cardiac Surgery: Identification of Biomarkers Associated With Postoperative Delirium. Frontiers in aging neuroscience 2021;13.
  45. 45.
    Kazmierski J, Banys A, Latek J, Bourke J, Jaszewski R. Raised IL-2 and TNF-α concentrations are associated with postoperative delirium in patients undergoing coronary-artery bypass graft surgery. International psychogeriatrics 2014;26:845– 855.
    OpenUrlCrossRefPubMed
  46. 46.
    Kazmierski J, Banys A, Latek J, Bourke J, Jaszewski R. Cortisol levels and neuropsychiatric diagnosis as markers of postoperative delirium:a prospective cohort study. Critical Care 2013;17:R38.
  47. 47.
    Lammers-Lietz F, Akyuz L, Feinkohl I, Lachmann C, Pischon T, Volk H-D, et al. Interleukin 8 in postoperative delirium – Preliminary findings from two studies. Brain, Behavior, & Immunity - Health 2022;20:100419.
  48. 48.
    Marcantonio ER, Rudolph JL, Culley D, Crosby G, Alsop D, Inouye SK. Serum biomarkers for delirium. The journals of gerontology Series A, Biological sciences and medical sciences 2006;61:1281–1286.
    OpenUrlCrossRefPubMedWeb of Science
  49. 49.↵
    Nurcahyo WI, Arifin A, Primatika AD, Muttaqin Z, Boom CE, Harahap MS, et al. An Association Between C-Reactive Protein Levels and the Occurrence of Cognitive Dysfunction After Heart Valve Replacement. Vascular Health and Risk Management 2021;17:713.
Back to top
Posted June 05, 2023.
Download PDF

Supplementary Material

Data/Code
Email
Multivariable models of postoperative delirium in cardiac surgery patients: proteomic and demographic contributions
Gonçalves MCB, Khera T, Otu HH, Narayanan S, Dillon ST, Shanker A, Gu X, Jung Y, Ngo LH, Marcantonio ER, Libermann TA, Subramaniam B
medRxiv 2023.05.30.23289741; doi: https://doi.org/10.1101/2023.05.30.23289741
Twitter logo Facebook logo LinkedIn logo Mendeley logo
Citation Tools

Subject Area