Systemic Inflammation and Metabolic Changes After Cardiac Surgery and Postoperative Delirium Risk

Abstract

Introduction: Postoperative delirium (POD) remains a major complication in geriatric surgical care, with poorly understood molecular mechanisms. Emerging evidence links cardiac surgery to elevated markers of neurologic injury, even in cognitively intact individuals. While neuroinflammation is the prevailing model, a more detailed characterization of the systemic inflammatory and metabolic response to surgery may offer deeper insights into POD pathogenesis. Methods: We used the 7K SomaLogic proteomic platform to analyze preoperative and postoperative day-one serum samples from 78 patients undergoing cardiac surgery with cardiopulmonary bypass. We compared proteomic profiles within individuals (pre- vs. post-surgery) and between those who developed POD and those who did not. Functional analyses were performed to identify relevant biological pathways. A composite metabo-inflammatory score (MIF) was derived to quantify systemic derangement. We modeled the association between POD and age, sex, baseline cognition, and MIF score. Results: Cardiac surgery with CPB was associated with marked inflammatory responses across all subjects, including increased IL-6, CRP, and serum amyloid A. Compared to controls, POD cases showed greater metabo-inflammatory shifts from baseline (average logFC = 2.56, p < 0.001). Lower baseline cognitive scores (OR = 0.74, p = 0.019) and higher MIF scores (OR = 1.03, p = 0.013) were independently associated with increased POD risk. Conclusions: Cardiac surgery with CPB elicits a significant metabo-inflammatory response in all patients. However, those who develop POD exhibit disproportionately greater dysregulation.

Keywords: cardiac surgery; inflammation; metabolism; postoperative delirium; proteomics.

Figure 1

Study design, data acquisition and analyses. An age- and sex-matched cohort of 78 patients was selected from the parent MINDDS trial of 364 non-mechanically ventilated patients who underwent non-emergent cardiac surgery with cardiopulmonary bypass. The participants were randomized to receive either dexmedetomidine or placebo and were followed subsequently with twice-daily assessment for delirium using the confusion assessment method. Sera for each participant in this nested study was comprehensively profiled for proteomic signatures using the high-throughput 7K SOMAScan proteomic platform. Comparative proteomic analyses between delirium cases and non-delirium controls were performed for differentially abundant proteins, among other analyses. CAM: confusion assessment method; CPB: cardiopulmonary bypass; POD+: postoperative delirium case; POD−: non-POD control.

Figure 2

(A) Principal component analyses (PCA) of all baseline and postoperative day-one samples, based on 202 protein features identified by cross-validated penalized logistic regression (Elastic Net) modeling. (B) Volcano plot of p-value (−log₁₀ scale) vs. fold-change (log₂ scale) of the 29 differentially abundant proteins, showing a four-fold postoperative change in select proteins, notably SAA1, SAA2, IL1RL1 and TNNI3, relative to baseline levels. (C) PCA of all samples at baseline, annotated by case–control status. (D) PCA of all samples on postoperative day one, annotated by case–control status. (E) Volcano plot of the 33 differentially abundant proteins among POD cases relative to controls at postoperative day one, which demonstrates significant changes in many markers of acute inflammation in addition to enzymes involved in energy metabolism. (F) Bar plot depicting the fold changes on postoperative day one relative to baseline for the 30 most differentially abundant proteins between POD cases and controls. Only the top 30 proteins were selected to facilitate clear visualization. n(Regulated): the number of proteins that reach statistical significance and the fold-change cut off for a given analysis. A4GNT: Alpha-1,4-N-acetylglucosaminyltransferase; ACHE: Acetylcholinesterase; ACTN2: Alpha-actinin-2; ALOX15B: Polyunsaturated fatty acid lipoxygenase ALOX15B; APOA5: Apolipoprotein A-V; ARRDC3: Arrestin domain-containing protein 3; BCAT2: Branched-chain-amino-acid aminotransferase, mitochondrial; CCL23: C-C motif chemokine 23; CHI3L1: Chitinase-3-like protein 1; CKB: Creatine kinase B-type; CKM: Creatine kinase M-type; CKMT2: Creatine kinase S-type, mitochondrial; CPB1: Carboxypeptidase B; CRP: C-reactive protein; DEFB106A: Beta-defensin 106; FCGR2A: Immunoglobulin gamma Fc region receptor II-a; FGFBP1: Fibroblast growth factor-binding protein 1; FGL1: Fibrinogen-like protein 1; GABARAPL1: Gamma-aminobutyric acid receptor-associated protein-like 1; GDF15: Growth/differentiation factor 15; GFAP: Glial fibrillary acidic protein; IGFBP2: Insulin-like growth factor-binding protein 2; IL1RL1: Interleukin-1 receptor-like 1; IL6: Interleukin-6; IZUMO4: Izumo sperm-egg fusion protein 4; LECT2: Leukocyte cell-derived chemotaxin-2; MDH1: Malate dehydrogenase, cytoplasmic; MYBPC3: Myosin-binding protein C, cardiac-type; NDUFA2: NADH dehydrogenase [ubiquinone] 1 alpha subcomplex subunit 2; NID2: Nidogen-2; NOG: Noggin; NQO2: Ribosyldihydronicotinamide dehydrogenase [quinone]; NTN1: Netrin-1; PBX1: Pre-B-cell leukemia transcription factor 1; PDE5A: cGMP-specific 3′,5′-cyclic phosphodiesterase; PGM5: Phosphoglucomutase-like protein 5; PML: Protein PML; POLR2J2: DNA-directed RNA polymerase II subunit RPB11-b1; PRL: Prolactin; SAA1: Serum amyloid A-1 protein; SAA2: Serum amyloid A-2 protein; SEMA5A: Semaphorin-5A; SFRP5: Secreted frizzled-related protein 5; SLC6A16: Orphan sodium- and chloride-dependent neurotransmitter transporter NTT5; TNC: Tenascin; TNNI2: Troponin I, fast skeletal muscle; TNNI3: Troponin I, cardiac muscle; TNNT2: Troponin T, cardiac muscle; TPI1: Triosephosphate isomerase; TTC36: Tetratricopeptide repeat protein 36; UBE2W: Ubiquitin-conjugating enzyme E2 W.

Figure 3

Functional analysis on protein signatures of CBP exposure and of delirium, for overrepresented (A) molecular functions, (B) biological processes, (C) subcellular components and (D) KEGG pathways.