Optimization of the cardiac delirium index by including age, decrease in butyrylcholinesterase actitivity, preoperative HbA1c, and postoperative hemoglobin levels: results of a secondary analysis of a prospective observational study

Abstract

Introduction: Postoperative delirium (POD) after cardiac surgery significantly affects the perioperative morbidity and mortality. Butyrylcholinesterase (BChE) is an enzyme primarily produced in the liver, which plays a crucial role in the hydrolysis of acetylcholine outside of neuronal synapses, referred to as extraneuronal hydrolysis. The integration of BChE activity into the cardiac delirium (CARDEL) index might increase its predictive power for identifying POD after cardiac surgery. Therefore, the primary aim of this study was to assess the applicability of the CARDEL index and determine whether integrating the BChE activity enables optimization of the predictive model.

Methods: This secondary analysis of a prospective observational study included patients undergoing elective coronary artery bypass graft surgery. BChE activity is expressed in units per liter (U/L), while the BChE drop refers to the percentage decrease in BChE activity from pre- to postoperative levels. POD risk factors were identified using multivariate regression analysis. The predictive power of the CARDEL index and an optimized model including BChE was calculated with receiver operating characteristic (ROC) analysis.

Results: Of 93 included patients, 20 (21.5%) developed POD. Elevated preoperative HbA1c [OR 2.5 (1.2-4.8), p = 0.01], a decrease in BChE activity [%, OR 1.1 (1.0-1.2), p = 0.04], age [1 (0.94-1.1), p = 0.55], and a postoperative hemoglobin change [OR 0.86 (0.78-0.96), p < 0.001] were identified as independent risk factors for POD. While the CARDEL index showed a moderate prediction of POD [AUCROC of 0.74 (0.60-0.87)], the optimization including BChE resulted in a significant prognostic improvement: AUCROC of 0.84 (0.72-0.94, p < 0.001).

Conclusion: Despite the small size of this derivation cohort, this study identified elevated HbA1c as the strongest risk factor for the development of POD, followed by a decrease in BChE activity, postoperative anemia, and age, respectively. By including these parameters to the CARDEL index, its predictive power for the identification of POD significantly improved in this derivation cohort. Moving forward, integrating these findings into clinical practice could enhance early risk stratification and targeted intervention for patients at high risk of POD. Therefore, further research should evaluate these results in a larger, external cohort.

Keywords: biomarker; cardiac artery bypass graft surgery; cardiac surgery; cholinergic system; postoperative delirium.

Figure 1

Boxplot demonstrating the differences in indices between both study groups calculated as (A) CARDEL index, (B) CARDEL index with drop in BChE activity included, and (C) index calculated from the optimized model. Box and whisker plots indicate median, interquartile range (box), minimum and maximum (whiskers). ***t-test p-value <0.001, **t-test p-value <0.01.

Figure 2

Spiderweb plot demonstrating the comparison of parameters resulting from univariate logistic regression analysis and those used in the CARDEL index calculation. *p < 0.05, **p < 0.01. Hb, hemoglobin; WBC, white blood cells; BChE, butyrylcholinesterase; HbA1c, glycated hemoglobin; POD, postoperative delirium; PWR, platelet-to-white-blood-cell ratio.

Figure 3

ROC for the CARDEL index (dotted grey line), the extended CARDEL index (dotted dark grey line), and the index according to the optimized model (solid black line). BChE, butyrylcholinesterase; CARDEL, cardiac delirium index; ROC, receiving operator curve.