Background

Impaired cognition is a major predisposing factor for postoperative delirium, but it is not systematically assessed. Anesthesia and surgery may cause postoperative delirium by affecting brain integrity. Neurofilament light in serum reflects axonal injury. Studies evaluating the perioperative course of neurofilament light in cardiac surgery have shown conflicting results. The authors hypothesized that postoperative serum neurofilament light values would be higher in delirious patients, and that baseline concentrations would be correlated with patients’ cognitive status and would identify patients at risk of postoperative delirium.

Methods

This preplanned secondary analysis included 220 patients undergoing elective cardiac surgery with cardiopulmonary bypass. A preoperative cognitive z score was calculated after a neuropsychological evaluation. Quantification of serum neurofilament light was performed by the Simoa (Quanterix, USA) technique before anesthesia, 2 h after surgery, on postoperative days 1, 2, and 5. Postoperative delirium was assessed using the Confusion Assessment Method for Intensive Care Unit, the Confusion Assessment Method, and a chart review.

Results

A total of 65 of 220 (29.5%) patients developed postoperative delirium. Delirious patients were older (median [25th percentile, 75th percentile], 74 [64, 79] vs. 67 [59, 74] yr; P < 0.001) and had lower cognitive z scores (–0.52 ± 1.14 vs. 0.21 ± 0.84; P < 0.001). Postoperative neurofilament light concentrations increased in all patients up to day 5, but did not predict delirium when preoperative concentrations were considered. Baseline neurofilament light values were significantly higher in patients who experienced delirium. They were influenced by age, cognitive z score, renal function, and history of diabetes mellitus. Baselines values were significantly correlated with cognitive z scores (r, 0.49; P < 0.001) and were independently associated with delirium whenever the patient’s cognitive status was not considered (hazard ratio, 3.34 [95% CI, 1.07 to 10.4]).

Conclusions

Cardiac surgery is associated with axonal injury, because neurofilament light concentrations increased postoperatively in all patients. However, only baseline neurofilament light values predicted postoperative delirium. Baseline concentrations were correlated with poorer cognitive scores, and they independently predicted postoperative delirium whenever patient’s cognitive status was undetermined.

Editor’s Perspective
What We Already Know about This Topic
  • Postoperative delirium is a common complication after cardiac surgery

  • Preoperative cognitive impairment is a leading predisposing factor of postoperative delirium

  • Elevated concentrations of baseline neurofilament light in serum have been associated with poorer baseline cognition in cardiac surgery patients.

What This Article Tells Us That Is New
  • In this preplanned secondary analysis of a prospective cohort study including 220 elective cardiac surgery patients, baseline serum neurofilament light concentrations were associated with the occurrence of postoperative delirium

  • Baseline serum neurofilament light concentrations were correlated with lower baseline cognitive scores and were an independent predictor of postoperative delirium when patients’ cognitive status was not considered

Postoperative delirium is a common complication with an incidence that varies between 20 and 45%, depending on the type of surgery and the studied population.1,2  It is associated with increased morbidity and mortality.3,4  Many predisposing and perioperative precipitating factors play an important role, among which advanced age is the best known.5–9  However, more important than the patient’s chronological age, it is the patient’s cognitive status that appears to be the leading cause of cognitive changes after anesthesia and surgery.10  Evaluating cognitive functions using a complete battery of neurocognitive tests is the accepted standard, but it requires expertise and manpower and is time-consuming in a preoperative setting. Therefore, other tools are needed to identify patients at risk of postoperative delirium.

Although the pathophysiology of delirium has not been completely elucidated, one plausible mechanism could be a direct effect of anesthesia and/or surgery on brain integrity.11–13  Similar to acute traumatic brain injury,14  central nervous system proteins are released into the systemic circulation during surgery, where they can be measured as a biomarker of neuronal injury.15  Although the ideal biomarker of neuronal injury is difficult to find, interest has grown for perioperative measurements of neurofilament light, a subunit of axonal neurofilaments. In this perspective, recent studies including nonintracranial, noncardiac surgical patients associated a greater rise of neurofilament light in serum after surgery with the occurrence and severity of postoperative delirium.16,17  Otherwise, delirium might also reveal some degree of brain frailty unmasked by a stressful surgical circumstance. In this regard, abnormally high baseline concentrations of neurofilament light in cerebrospinal fluid or serum have been demonstrated to indicate ongoing neurodegeneration16,18  and were associated with the occurrence of postoperative delirium in noncardiac surgery.17,19 

There are only few studies in cardiac surgery evaluating perioperative evolution of neurofilament light, and most of them include a small number of patients.20–25  Moreover, they show conflicting results, which may be partially due to heterogeneous timings of analyses and, for some of them, lack of postoperative follow-up. Among these studies, one case series including nine patients correlated postoperative serum neurofilament light and the occurrence of postoperative delirium.22  However, no association was found with the preoperative values for this type of surgery.

The aim of this prospective preplanned secondary analysis is to bring insights into the perioperative course of serum neurofilament light in a large, well-characterized cohort of cardiac surgery patients. On one hand, we hypothesized that postoperative values of neurofilament light would be higher in patients who developed postoperative delirium. On the other hand, we sought to clarify whether baseline concentrations would be correlated with patient’s preoperative cognitive status, and thus would be a reliable alternative in identifying patients at risk of this complication.

Population and Study Design

This preplanned secondary study is part of a monocentric prospective observational project in which the primary outcome is to assess the association between lower intraoperative frontal α-band power of the electroencephalogram (EEG) and the occurrence of postoperative delirium.26  This project was approved on September 2018 by Hospital and Faculty Ethics Committee - Saint-Luc University Hospital (2018/20SEP/350; Brussels, Belgium; chairman, J.-M. Maloteaux) and registered in ClinicalTrials.gov (NCT03706989; principal investigator, Mona Momeni, M.D., Ph.D.; date of registration, October 11, 2018) before the start of the trial. Written informed consent was obtained from all patients according to the Declaration of Helsinki. Enrollment started on May 15, 2019, and was completed on December 15, 2021.

The cohort included adult patients undergoing a first elective cardiac surgery with a sternotomy approach and using a normothermic cardiopulmonary bypass (CPB). Surgical exclusion criteria were emergencies, reinterventions, endocarditis, ventricular assist devices, heart transplantation, and mini-invasive and robotic cardiac surgery. Personal/clinical exclusion criteria were preoperative delirium, psychiatric disorder, non–French-speaking patients, preoperative renal replacement therapy, chronic alcoholism, preoperative liver dysfunction (liver function tests threefold the upper normal value), and use of antiepileptic medication.

Preoperative Neuropsychological Testing

A battery of five neuropsychological tests, similar to the work of McDonagh et al.,27  was performed the day before the surgery. It consisted of (1) the 16-item Free and Cued Selective Reminding Test, (2) the Modified Visual Reproduction Test from the Wechsler Memory Scale, (3) the Digit Span Test from the revised version of the Wechsler Adult Intelligence Scale, (4) the Trail Making Test, and (5) the Digit Symbol test from the Wechsler Adult Intelligence Scale.

To score the patients’ preoperative cognitive status, sample-specific z scores [(individual result – mean of the studied population)/SD of the studied population] were computed from five different results: (1) the sum of the results of the three free recalls of the Free and Cued Selective Reminding Test, (2) the result of the Modified Visual Reproduction Test, (3) the result of the Digit Span Test, (4) the result of the Trail Making Test (time part B – time part A), and (5) the result of the Digit Symbol Test. Ultimately, a global cognitive z score was calculated by averaging the five z scores.

To limit any bias, only two persons trained by neuropsychologists from our Department of Neurology performed the preoperative neurocognitive evaluation.

Anesthesia and CPB protocols

Anesthesia and CPB protocols were standardized to limit biases regarding the risk of developing postoperative delirium.

Premedication consisted of alprazolam (0.25 to 0.5 mg). Besides the standard monitoring (pulse oximetry, 12-lead electrocardiogram, invasive blood pressure in radial or femoral artery, central venous pressure via an internal jugular vein catheter), a Neuro-SENSE depth-of-anesthesia monitor (NeuroWave Systems Inc., USA) and a bilateral cerebral oximeter (INVOS 5100, Somanetics Corp., USA) were systematically used in accordance to institutional protocol. Anesthesia was induced with midazolam (0.03 to 0.06 mg · kg-1), sufentanil (10 to 50 μg, titrated) and propofol, according to the raw frontal EEG and spectral analysis provided by the Neuro-SENSE monitor. Either cisatracurium (0.2 mg/kg) or rocuronium (0.5 mg/kg) was used for muscle relaxation. Anesthesia was maintained with sevoflurane and a continuous infusion of sufentanil (0.4 to 0.5 μg · kg–1 · h–1) and guided by the spectral analysis provided by Neuro-SENSE monitor. Intraoperative administration of ketamine was not part of the study protocol. Boluses of ephedrine (5 mg) and/or a continuous infusion of norepinephrine were used to maintain the mean arterial blood pressure around 60 to 70 mmHg. According to institutional protocol, an intraoperative algorithm was used to optimize cerebral oxygenation and to avoid any intraoperative EEG suppression.

The management of the CPB occurred under the supervision of the cardiac anesthesiologist. The protocol consisted in a normothermic system with continuous nonpulsatile flow of 2.4 l · min–1 · m–2, with a 6 or 8 liter per minute oxygenator module (Inspire LivaNova, Sorin Group, Italy). The pump was primed with 1,000 ml acetated Ringer’s solution (Plasmalyte A, Baxter S.A., Belgium) and mannitol 15% 2 ml/kg (Baxter S.A., Belgium). Myocardial protection was provided every 15 min by warm blood enriched with potassium chloride and magnesium sulfate. The hematocrit was maintained between 25 and 35%. A pH-stat method was used to maintain pH and Paco2 within normal ranges.

After the procedure, patients were transferred to the cardiovascular intensive care unit (ICU) with a continuous infusion of propofol (1 to 3 mg · kg–1 · h–1). Postoperative analgesia consisted of a patient-controlled analgesia pump containing morphine or piritramide (2 mg/ml), paracetamol 1 g/6 h (for the first 2 days, then as pro re nata), and if no contraindication, ibuprofen 400 mg/8 h (for the first 2 days, then as pro re nata).

Perioperative Blood Samples

Neurofilament Light Quantification in Serum

Five blood samples were taken for neurofilament light measurements: (1) before anesthesia, (2) 2 h after the admission in the ICU, (3) at postoperative day 1, (4) at postoperative day 2, and (5) at postoperative day 5. Blood samples were collected in 4.9-ml serum tubes (S-Monovette, Sarstedt B.V., Germany). After centrifugation (1,800 rounds/min, room temperature, 10 min), the serum was aliquoted in 1.8-ml polypropylene microtubes (VWR, Belgium) and stored at –80°C in the biobank of Cliniques Universitaires Saint-Luc (Brussels, Belgium). Single-molecule array technology (Simoa, Quanterix, USA) was used for neurofilament light quantification.28  All the samples were sent on dry ice to the neurochemistry laboratory of the Amsterdam University Medical Centers (Amsterdam, The Netherlands). Measurements were performed in singlicate by certified technicians who were blinded to clinical information. Interassay coefficients of variation were 11.9% for a quality control sample with a neurofilament light concentration of 16.4 pg/ml and 8.8% for a quality control sample with a concentration of 170.3 pg/ml. Regarding the repeatability, intra-assay coefficients of variation were 4.3% for a neurofilament light concentration of 16.4 pg/ml and 2.6% for a concentration of 170.3 pg/ml. Both inter- and intra-assay coefficients of variation were defined according to the work of Andreasson et al.29  The functional lower limit of quantification of neurofilament light was 0.7 pg/ml. All analyses were supervised by the head of the Department of Clinical Chemistry in Amsterdam (C.E.T.). We did not perform blood samples at postoperative day 5 for the first 10 included patients as we decided to add this fifth sample once the study had initiated.

Apolipoprotein E Genotyping.

Apolipoprotein E genotyping was part of the study protocol. Whole blood was collected in 3.4-ml EDTA tubes (S-Monovette, Sarstedt B.V., Germany) and stored at –4°C for a maximum of 48 h before analysis by Institut de Pathologie and Génétique (Gosselies, Belgium). Apolipoprotein E haplotype status was determined by polymerase chain reaction amplification, followed by Sanger sequencing. Patients were considered ε4(+) when they carried at least one allele ε4. It was hypothesized that the presence of ε4 would directly influence our results.

Postoperative Delirium Screening

In the ICU, all the patients were kept sedated with a continuous infusion of propofol until they reached several criteria for weaning from ventilation, according to local standard protocols. Participants were screened for postoperative delirium using the Confusion Assessment Method for Intensive Care Unit30  after an evaluation of the level of sedation by the Richmond Agitation-Sedation Scale (greater than –3).31  The screening was performed by trained nurses three times a day until discharge from the ICU. At the ward, the Confusion Assessment Method was performed two times a day by trained nurses until the patient’s discharge from the hospital.32  The training of the nursing staff, both in the ICU and at the ward, was initiated 2 yr before the start of this study as part of the hospital accreditation. Considering the fluctuating course of postoperative delirium, a flowchart review was systematically associated with these tools to improve their accuracy.33  The medical chart of the patients was checked by the research staff for any notifications made by nurses or physicians suggesting an episode of postoperative delirium (e.g., aggressive or inappropriate behavior, confusion, use of restraints, use of haloperidol, reports of hallucinations).

Statistical Analyses

This is a secondary preplanned analysis of a prospective project in cardiac surgery patients. The sample size was based on the primary objective of the study that evaluated whether the intraoperative frontal EEG α-band power is associated with the occurrence of postoperative delirium. A minimum of 204 patients were needed to answer our primary hypothesis. We included 220 patients to take into consideration any dropouts.

The Kolmogorov–Smirnov test was used to check the normality of the data. Categorical data are presented as numbers and percentages. Continuous variables are presented as mean ± SD or median [25th percentile, 75th percentile], depending on whether they were normally distributed or not. Comparison of continuous variables between patients with or without postoperative delirium was performed with an independent t test or Mann–Whitney U test, depending on the normality assumption. A Pearson chi-square test or Fisher exact test was used to compare categorical variables between the two groups.

A linear mixed-effects model was used with postoperative neurofilament light concentrations (log10) as outcome variable, delirium and time as fixed effects, and patients as random effects. Time was considered as a within-patient factor in mixed models because of the repeated measures design.

A univariate linear regression analysis was performed to evaluate whether the severity of cardiac surgery (estimated by surgical time, CPB time, and peak of postoperative troponin in serum) had a significant influence on the highest value of postoperative neurofilament light (log10) available.

A Pearson correlation coefficient was used to assess correlation between global cognitive z score and baseline neurofilament light concentrations (log10).

A univariate linear regression analysis sought the relationship between baseline neurofilament light (log10) and different covariates known from the literature that influence its values (age, sex, body mass index, apolipoprotein Eε4 carrier status, global cognitive z score, glomerular filtration rate, and history of diabetes mellitus).34–37  The relationship between baseline neurofilament light (log10) and the risk of postoperative delirium was explored in Cox proportional hazard regression models by entering baseline neurofilament light and covariates that significantly influenced its value in univariate analysis. The time to event (postoperative delirium) was censored at postoperative day 10 as most of our patients leave the hospital before day 10, and as postoperative delirium mostly occurred in the first three days postoperatively. The proportional hazard assumption was tested by examination of Kaplan–Meier curves. A P value less than 0.05 was considered statistically significant. All statistical analyses were performed using IBM SPSS Statistics version 27 (Macintosh, USA).

The flowchart of the study is detailed in figure 1. In total, 220 patients completed the study. There were no missing data regarding the reporting of delirium episodes during patients’ hospital stay. The incidence of postoperative delirium was 29.5%. Among the 65 delirious patients, 34 cases were categorized as “hypoactive” (52.3%), 18 as “hyperactive” (27.7%), and 13 as “mixed” (20.0%) subtypes. Thirty-five patients (53.8%) incurred their first episode of delirium on postoperative day 1 (see figure, Supplemental Digital Content 1, https://links.lww.com/ALN/D452, detailing the cumulative incidence of the first episode of postoperative delirium).

Fig. 1.

Flowchart of the study.

Fig. 1.

Flowchart of the study.

Close modal

Preoperative patients’ demographics and comorbidities are listed in table 1. Patients who experienced postoperative delirium were significantly older (74 [64, 79] vs. 67 [59, 74] yr; P < 0.001). They also had higher European System for Cardiac Operative Risk Evaluation (EuroSCORE) II scores (2.4 [1.3, 4.0] vs. 1.5 [0.9, 2.5] %; P < 0.001), lower global cognitive z score (–0.52 ± 1.14 vs. 0.21 ± 0.84; P < 0.001), and lower preoperative glomerular filtration rate (70 [55, 86] vs. 79 [63, 90] ml/min; P = 0.016).

Table 1.

Preoperative Characteristics and Comorbidities of Patients with and without Postoperative Delirium

Preoperative Characteristics and Comorbidities of Patients with and without Postoperative Delirium
Preoperative Characteristics and Comorbidities of Patients with and without Postoperative Delirium

Table 2 demonstrates intraoperative and postoperative data. CPB time was significantly longer in delirious patients (112 ± 40 vs. 99 ± 33 min; P = 0.015), but no differences were to be noted regarding the surgical time and the type of surgery. Delirious patients stayed significantly longer in the ICU (3.5 ± 3 vs. 2 ± 1 days; P < 0.001) and in the hospital (8 [7, 10.5] vs. 8 [7, 9] days; P = 0.006).

Table 2.

Intraoperative and Postoperative Data of Patients with and without Postoperative Delirium

Intraoperative and Postoperative Data of Patients with and without Postoperative Delirium
Intraoperative and Postoperative Data of Patients with and without Postoperative Delirium

Pre- and postoperative values of serum neurofilament light (pg/ml) in patients with and without postoperative delirium are presented in figure 2 (see table, Supplemental Digital Content 2, for concentrations details, https://links.lww.com/ALN/D453). One sample at 2 h after surgery was excluded for quality reasons (repeated clotting of the sample during the quantification process), and six patients either withdrew their agreement for the fifth blood sample or were transferred to another hospital before postoperative day 5. Eventually, a total of 1,083 blood samples was analyzed. Neurofilament light concentrations increased after surgery for all the patients, with the highest concentration reached at postoperative day 5 for 199 patients (94.6% of the cases in whom concentrations on day 5 were available). Pre- and postoperative neurofilament light values were significantly higher at all time points in patients who experienced postoperative delirium (fig. 2). The time trends of postoperative neurofilament light values of patients with and without postoperative delirium presented a similar course from 2 h after the admission in the ICU (first postoperative sample) to postoperative day 5 (last postoperative sample; fig. 3A). However, when considering normalized postoperative neurofilament light values (log10) to preoperative values (log10), the time trends overlapped, confirming that the two groups differed mainly by their baseline values (fig. 3B). Medians and percentiles of postoperative neurofilament light values after subtracting baseline concentrations, stratified by delirium status, are detailed in the table in Supplemental Digital Content 3 (https://links.lww.com/ALN/D454).

Fig. 2.

Perioperative concentrations of neurofilament light in serum (pg/ml) in patients with and without postoperative delirium. Neurofilament light concentrations are represented as medians and 95% CI error bars. Numbers of patients with available data are mentioned between brackets. Postoperative timings for neurofilament light measurement are expressed in hours after surgery.

Fig. 2.

Perioperative concentrations of neurofilament light in serum (pg/ml) in patients with and without postoperative delirium. Neurofilament light concentrations are represented as medians and 95% CI error bars. Numbers of patients with available data are mentioned between brackets. Postoperative timings for neurofilament light measurement are expressed in hours after surgery.

Close modal
Fig. 3.

Time trends of perioperative serum neurofilament light concentrations in patients with and without postoperative delirium. (A) The slopes represent the postoperative release of neurofilament light in serum in patients with (red line) and without (blue line) postoperative delirium. Neurofilament light concentrations are expressed as geometric mean ± 95% CI (pg/ml). (B) Perioperative time trends of neurofilament light concentrations, from preoperative sample to postoperative day 5, stratified by delirium status (red line, delirious patients; blue line, nondelirious patients). Neurofilament light values were normalized by log-transforming pre- and postoperative values, and then by subtracting the log-transformed baseline values from the log-transformed postoperative values. Postoperative timings for neurofilament light measurement are expressed in hours after surgery.

Fig. 3.

Time trends of perioperative serum neurofilament light concentrations in patients with and without postoperative delirium. (A) The slopes represent the postoperative release of neurofilament light in serum in patients with (red line) and without (blue line) postoperative delirium. Neurofilament light concentrations are expressed as geometric mean ± 95% CI (pg/ml). (B) Perioperative time trends of neurofilament light concentrations, from preoperative sample to postoperative day 5, stratified by delirium status (red line, delirious patients; blue line, nondelirious patients). Neurofilament light values were normalized by log-transforming pre- and postoperative values, and then by subtracting the log-transformed baseline values from the log-transformed postoperative values. Postoperative timings for neurofilament light measurement are expressed in hours after surgery.

Close modal

The linear mixed-effects model analyzing postoperative neurofilament light showed a significant main effect for delirium status (P < 0.001; β, –0.15 [95% CI, –0.21 to –0.08]) and a significant main effect for time (P < 0.001; β, 0.10 [95% CI, 0.09 to 0.10]). There was no significant interaction between delirium and time (P = 0.158).

A univariate linear regression analysis was used to correlate the highest value of postoperative neurofilament light concentration (log10) in serum with, respectively, the surgical time, CPB time, and peak of postoperative troponin values. These correlations were poor (all Pearson correlation r < 0.21) as demonstrated in the table in Supplemental Digital Content 4 (https://links.lww.com/ALN/D455).

Univariate linear regression analysis revealed that older age, lower global cognitive z score, lower baseline glomerular filtration rate, and the presence of diabetes mellitus were associated with higher baseline neurofilament light concentrations (log10; table 3). Otherwise, baseline neurofilament light values were significantly correlated with preoperative global cognitive z scores (Pearson correlation r, 0.49; P < 0.001; fig. 4). A Cox proportional hazard regression analysis showed that among baseline neurofilament light, age, baseline glomerular filtration rate, history of diabetes mellitus, and global cognitive z score, only a poorer global cognitive z score significantly increased the hazard of developing postoperative delirium (P = 0.001; hazard ratio, 0.62 [95% CI, 0.46 to 0.83]; table 4A). Considering the strong association between global cognitive z score and baseline neurofilament light values (fig. 4), we performed a Cox proportional hazard regression analysis excluding the global cognitive z score. This second Cox regression analysis would represent a clinical situation where the patient’s preoperative cognitive status is undetermined. This analysis showed that higher neurofilament light concentrations (log10) at baseline increased the hazard of developing postoperative delirium (P = 0.037; hazard ratio, 3.34 [95% CI, 1.07 to 10.4]), independently from age and other covariates (table 4B). Moreover, the results were similar using EuroSCORE II as a proxy for comorbidities’ load. They confirmed that whenever cognitive status is not evaluated before cardiac surgery, quantifying baseline serum neurofilament light might help identify vulnerable patients better than considering cardiovascular comorbidities (table, Supplemental Digital Content 5, https://links.lww.com/ALN/D456).

Table 3.

Univariate Linear Regression Analysis—Dependent Variable: Serum Neurofilament Light (log10) at Baseline

Univariate Linear Regression Analysis—Dependent Variable: Serum Neurofilament Light (log10) at Baseline
Univariate Linear Regression Analysis—Dependent Variable: Serum Neurofilament Light (log10) at Baseline
Table 4.

Cox Proportional Hazard Regression Analysis—Dependent Variable: Postoperative Delirium

Cox Proportional Hazard Regression Analysis—Dependent Variable: Postoperative Delirium
Cox Proportional Hazard Regression Analysis—Dependent Variable: Postoperative Delirium
Fig. 4.

Correlation between preoperative global cognitive z score and baseline neurofilament light concentrations (log10) in serum.

Fig. 4.

Correlation between preoperative global cognitive z score and baseline neurofilament light concentrations (log10) in serum.

Close modal

The results of our study provide evidence that cardiac surgery is associated with axonal injury. Indeed, we demonstrated that serum neurofilament light concentrations increased gradually after surgery in all subjects, up to postoperative day 5. These results are in line with other recent studies. Barbu et al.24  showed a postoperative rise of serum neurofilament light in a cohort of 61 patients who underwent uncomplicated cardiac surgery with CPB. Similarly, Wiberg et al.25  described the same increase in a cohort of 168 cardiac surgery patients. Potential leading causes of neuronal damage in cardiac surgery are numerous, including inflammation, microembolic load, neurotoxicity, and oxidative stress, among many others.6,13  Moreover, a previous study showed that serum neurofilament light still increased up to 1 week after cardiac surgery.23  The mechanisms underlying this slow temporal course of release in serum after surgery remain unclear and might be partially due to a passive liberation of subunits of injured neurofilaments into the circulation, rather than an active process.23  In fact, little is known about the neurofilament light kinetics and transport from the brain to the blood after an acute brain insult. Cardiac surgery is associated with blood–brain barrier disruption, and this phenomenon has been suggested as a potential causal mechanism for postoperative neurocognitive disorders.6,38,39  However, whether this increased blood–brain barrier permeability could explain the rise in serum neurofilament light concentrations is unclear. For instance, a recent study by Kalm et al.40  suggested that neurofilament light release in serum was not related to blood–brain barrier permeability. To date, neither the metabolization of neurofilament light nor its clearance from the circulation have been completely defined, although studies have shown that concentrations in blood might be partially affected by renal function, especially in older adults.34,35  The type of surgery might also have an influence. Indeed, studies suggest that cardiac surgery with CPB, by worsening the severity of neuronal damage,41  might be partially responsible for a more prolonged presence of high concentrations of neurofilament light in serum.21–23  Surprisingly, we found only poor correlations between markers of cardiac surgery severity (surgical time, CPB time, peak of postoperative troponin in serum) and the highest postoperative neurofilament light value. A plausible explanation might come from the fact that we were not able to determine the actual peak of postoperative neurofilament light concentration because we stopped the quantification at day 5, whereas the concentrations were possibly still increasing.

This study also confirmed that postoperative serum neurofilament light values were significantly higher at all time points in delirious patients. These findings are in line with the results of the case series by Saller et al.,22  as well as other studies, in noncardiac surgery,16,17  which also found an association between postoperative neurofilament light concentrations and the occurrence of postoperative delirium.16,17  However, in contradiction with the study by Casey et al.,16  this difference between both groups disappeared for the first 2 days postoperatively when baseline neurofilament light values were taken into account. Some methodologic parameters might explain the differences between the two studies, such as the matrix used for quantification (plasma vs. serum) or variabilities in assays analytical performances. Nonetheless, the absence of difference in the rates of postoperative release of serum neurofilament light between both groups in the early postoperative period questions the predictability of postoperative serum neurofilament light for early postoperative neurocognitive disorders, such as delirium. First, our findings do not support the “brain failure” hypothesis,12  as later during the hospitalization, nondelirious patents reached values similar to those of delirious patients, but without any symptoms. Second, the kinetics of postoperative serum neurofilament light release did not correspond to clinical symptomatology, as most of our patients returned to a nondelirious clinical state, although serum neurofilament light concentrations were still increasing. One could hypothesize that postoperative rise of neurofilament light in serum might also be related to peripheral nerve injuries during surgery. Indeed, in a recent study in orthopedic surgery, Danielson et al. did not observe significant cerebrospinal fluid changes in neurofilament light concentrations after surgery, whereas there was an early and progressive increase in neurofilament light concentrations in serum.42  Otherwise, another study in nonintracranial, nonartery carotid surgery recently found an association between the peak of postoperative neurofilament light in serum and postoperative covert stroke, a complication that was not investigated in our study.43  To summarize, postoperative release of neurofilament light in serum after cardiac surgery is probably multifactorial, and future studies are needed to determine whether this gradual increase is due to a slow clearance after surgery-induced injuries or related to ongoing neuronal injury, which might be responsible for the occurrence of postoperative neurocognitive disorders.

Our results moreover demonstrated that patients who developed postoperative delirium had significantly higher baseline neurofilament light concentrations. Baseline neurofilament light concentrations have been previously associated with preoperative white matter injury and reduced hippocampal volume, and hence could be a serum marker of reduced brain integrity.16  Postoperative delirium might thus be the consequence of underlying neurodegeneration and unmasked by a stressful surgical event. The association between higher baseline neurofilament light concentrations and postoperative delirium is still controversial.16,17,19,22  Nonetheless, our results suggest that patients with pre-existing, even subclinical neurodegeneration might be more susceptible to develop delirium and that further studies are needed to corroborate this hypothesis. We also showed that among baseline neurofilament light concentration, age, baseline glomerular filtration rate, history of diabetes mellitus, and global cognitive z score, the last was the only factor that significantly and independently of others increased the risk of developing postoperative delirium. Indeed, pre-existing cognitive impairment is among its strongest predisposing factors.10  However, preoperative screening for cognitive impairment is not systematically performed as it is time-consuming in a busy surgical setting. Therefore, to model a clinical situation in which cognitive status is not evaluated before surgery, we performed a Cox regression analysis excluding cognitive results, and we found that baseline serum neurofilament light significantly increased the hazard of developing postoperative delirium. Moreover, we found a strong correlation between the global cognitive z score and baseline neurofilament light values. In the surgical setting, our results corroborate the recent findings of Brown et al.,44  who demonstrated that poorer preoperative cognitive scores were associated with higher values of neurofilament light in serum in patients undergoing cardiac surgery. Altogether, our findings suggest that baseline serum neurofilament light could be a reliable, clinically feasible biologic tool to identify patients at risk of postoperative delirium whenever the patient’s preoperative cognitive status cannot be assessed. Additionally, we found that the impact of baseline neurofilament light values on the hazard of developing postoperative delirium appeared regardless of age and renal function. This finding is important as age and renal function have both been shown to influence neurofilament light values.34,35,45  It moreover questions the importance of patient’s chronological age as a predisposing factor of postoperative delirium.

Otherwise, we did not find a direct influence of apolipoprotein Eε4 on the occurrence of postoperative delirium or on baseline neurofilament light values. These findings are in line with previous studies46,47  and suggest that apolipoprotein Eε4 genotyping might not be suitable for postoperative delirium prediction.

Limitations

A first limitation of this study is that we did not use a control group to evaluate preoperative cognitive functions. Instead, we used raw test scores’ distribution to generate norms from the mean and SD of the population to ultimately transform these raw scores into an easily interpretable score, a z score. There is a growing interest for the use of the Mini-Cog testing, which might represent a reliable and easy-to-use alternative to complete neuropsychological testing.48  Further studies evaluating the potential added value of measuring baseline neurofilament light values to preoperative Mini-Cog test results would be of interest for identification of patients at risk of postoperative delirium.

Second, we did not evaluate the relationship between neurofilament light concentrations and the severity or the duration of postoperative delirium. Indeed, this secondary preplanned study only focused on the occurrence of postoperative delirium.

Third, this study was not designed to investigate the inflammatory hypothesis of postoperative delirium and neurofilament light release in serum. Indeed, our protocol did not include inflammatory cytokines or biomarkers of blood–brain barrier integrity.

Fourth, concomitant use of other biomarkers such as tau, glial fibrillary acid protein, or ubiquitin C-terminal hydroxylase L1 would have been helpful to clarify the mechanisms of brain injury in the perioperative period. It should be noted that the potential association of these biomarkers with neurologic outcomes is currently under investigation in the surgical context.13,17,19,49,50 

Finally, these data cannot be extrapolated to all cardiac surgery patients. Further studies including a more exhaustive panel of cardiac pathologies and types of surgery might add some critical information regarding the perioperative release of neurofilament light in this specific population.

Conclusions

This prospective, preplanned secondary study demonstrated that cardiac surgery with CPB is associated with axonal injury as postoperative serum neurofilament light concentrations continued to increase in all patients. However, the trend of postoperative neurofilament light concentrations from baseline could not predict postoperative delirium in our study. Additionally, higher neurofilament light values at baseline were associated with poorer preoperative cognitive test results and independently increased the hazard of developing postoperative delirium in a model where the patient’s preoperative cognitive status was unknown. Baseline serum neurofilament light might therefore be an indicator of underlying neurocognitive impairment.

Acknowledgments

The authors would like to thank Bernard Grisart, Ph.D. (University of Liège, Liège, Belgium; Institute of Pathology and Genetics, Gosselies, Belgium) for performing Apolipoprotein E genotyping and helping to carry out this research project in time. The authors also would like to show their gratitude to their colleagues, Laetitia Miltoni, M.D., Clinical Research Coordinator (Department of Anesthesiology, Saint-Luc University Hospital, Brussels, Belgium), who assisted in data acquisition, and Arnaud Steyaert, M.D. (Department of Anesthesiology, Saint-Luc University Hospital, Brussels, Belgium) who assisted in the design of some figures. They also thank Eline Willemse, Ph.D. (Department of Clinical Chemistry, Amsterdam University Medical Centers, Amsterdam, The Netherlands), for her help with coordinating serum shipments. Finally, the authors gratefully acknowledge the contributions of the patients and the colleagues (perfusionists, nurses, laboratory technicians, anesthesiologists, surgeons, intensive care unit physicians) who took part in this study.

Research Support

This work was supported by the Belgian Society of Anesthesiology, Resuscitation, Perioperative Medicine and Pain Management (Evergem, Belgium) and by the Fund for Scientific Research (Brussels, Belgium).

Competing Interests

The current research of Dr. Teunissen is supported by the European Commission (Marie Curie International Training Network, grant agreement No. 860197, MIRIADE), Innovative Medicines Initiatives 3TR (Horizon 2020, grant No. 831434), European Platform for Neurodegenerative Diseases (IMI 2 Joint Undertaking, grant No. 101034344), EU Joint Program - Neurodegenerative Disease Research (bPRIDE), National MS Society (Progressive MS Alliance), Alzheimer Drug Discovery Foundation, Alzheimer Association, Health Holland, the Dutch Research Council (ZonMW), Alzheimer Drug Discovery Foundation, Selfridges Group Foundation, and Alzheimer Netherlands. Dr. Teunissen is the recipient of ABOARD, which is a public–private partnership receiving funding from ZonMW (No. 73305095007) and Health~Holland, Topsector Life Sciences & Health (PPP-allowance; No. LSHM20106). Dr. Teunissen is the recipient of TAP-dementia, a ZonMW-funded project (No. 10510032120003) in the context of the Dutch National Dementia Strategy. Dr. Teunissen performed contract research for ADx Neurosciences (Tokyo, Japan), AC-Immune (Lausanne, Switzerland), Aribio (Seongnam, South Korea), Axon Neurosciences (Bratislava, Slovakia), Beckman-Coulter (Brea, California), BioConnect (Huissen, The Netherlands), Bioorchestra (Daejeon, South Korea), Brainstorm Therapeutics (New York, New York), Celgene (Summit, New Jersey), Cognition Therapeutics (Purchase, New York), EIP Pharma (Boston, Massachusetts), Eisai (Tokyo, Japan), Eli Lilly Fujirebio (Tokyo, Japan), Grifols (Barcelona, Spain), Instant Nano Biosensors (Taipei City, Taiwan), Merck (Rahway, New Jersey), Novo Nordisk (Bagsvaerd, Denmark), Olink (Uppsala, Sweden), PeopleBio (Seongnam, South Korea), Quanterix (Boston, Massachusetts), Roche (Bâle, Switzerland), Siemens (Munich, Germany), Toyama (Toyama, Japan), and Vivoryon (Munich, Germany). The other authors declare no competing interests.

Supplemental Digital Content 1. Cumulative incidence of postoperative delirium, https://links.lww.com/ALN/D452

Supplemental Digital Content 2. Perioperative concentrations of neurofilament light in serum (pg/ml) in patients with and without postoperative delirium, https://links.lww.com/ALN/D453

Supplemental Digital Content 3. Postoperative neurofilament light concentrations (without baseline), https://links.lww.com/ALN/D454

Supplemental Digital Content 4. Univariate analysis – Severity of surgery and postoperative neurofilament light, https://links.lww.com/ALN/D455

Supplemental Digital Content 5. Cox proportional hazard regression analysis including EuroSCORE II, https://links.lww.com/ALN/D456

1.
Evered
L
,
Silbert
B
,
Knopman
DS
et al
;
Nomenclature Consensus Working Group
:
Recommendations for the nomenclature of cognitive change associated with anaesthesia and surgery – 2018.
Anesthesiology
2018
;
129
:
872
9
2.
Rudolph
JL
,
Marcantonio
ER
:
Review articles: Postoperative delirium: Acute change with long-term implications.
Anesth Analg
2011
;
112
:
1202
11
3.
Witlox
J
,
Eurelings
LS
,
de Jonghe
JF
,
Kalisvaart
KJ
,
Eikelenboom
P
,
van Gool
WA
:
Delirium in elderly patients and the risk of postdischarge mortality, institutionalization, and dementia: A meta-analysis.
JAMA
2010
;
304
:
443
51
4.
Gleason
LJ
,
Schmitt
EM
,
Kosar
CM
et al
:
Effect of delirium and other major complications on outcomes after elective surgery in older adults.
JAMA Surg
2015
;
150
:
1134
40
5.
Maldonado
JR
:
Delirium pathophysiology: An updated hypothesis of the etiology of acute brain failure.
Int J Geriatr Psychiatry
2018
;
33
:
1428
57
6.
Berger
M
,
Terrando
N
,
Smith
SK
,
Browndyke
JN
,
Newman
MF
,
Mathew
JP
:
Neurocognitive function after cardiac surgery: From phenotypes to mechanisms.
Anesthesiology
2018
;
129
:
829
51
7.
Aldecoa
C
,
Bettelli
G
,
Bilotta
F
et al
:
European Society of Anaesthesiology evidence-based and consensus-based guideline on postoperative delirium.
Eur J Anaesthesiol
2017
;
34
:
192
214
8.
Cereghetti
C
,
Siegemund
M
,
Schaedelin
S
et al
:
Independent predictors of the duration and overall burden of postoperative delirium after cardiac surgery in adults: An observational cohort study.
J Cardiothorac Vasc Anesth
2017
;
31
:
1966
73
9.
Chen
H
,
Mo
L
,
Hu
H
,
Ou
Y
,
Luo
J
:
Risk factors of postoperative delirium after cardiac surgery: A meta-analysis.
J Cardiothorac Surg
2021
;
16
:
113
10.
Evered
LA
:
Predicting delirium: Are we there yet?
Br J Anaesth
2017
;
119
:
281
3
11.
Newman
MF
,
Mathew
JP
,
Grocott
HP
et al
:
Central nervous system injury associated with cardiac surgery.
Lancet
2006
;
368
:
694
703
12.
Evered
L
,
Silbert
B
,
Scott
DA
,
Zetterberg
H
,
Blennow
K
:
Association of changes in plasma neurofilament light and tau levels with anesthesia and surgery. Results from the CAPACITY and the ARCADIAN studies.
JAMA Neurol
2018
;
75
:
542
7
13.
Lopez
MG
,
Hughes
CG
,
DeMatteo
A
et al
:
Intraoperative oxidative damage and delirium after cardiac surgery.
Anesthesiology
2020
;
132
:
551
61
14.
Czeiter
E
,
Amrein
K
,
Gravesteijn
BY
et al
;
CENTER-TBI Participants and Investigators
:
Blood biomarkers on admission in acute traumatic brain injury: Relations to severity, CT findings and care path in the CENTER-TBI study.
EBioMedicine
2020
;
56
:
102785
15.
Cata
JP
,
Abdelmalak
B
,
Farag
E
:
Neurological biomarkers in the perioperative period.
Br J Anaesth
2011
;
107
:
844
58
16.
Casey
CP
,
Lindroth
H
,
Mohanty
R
et al
:
Postoperative delirium is associated with increased plasma neurofilament light.
Brain
2020
;
143
:
47
54
17.
Fong
TG
,
Vasunilashorn
SM
,
Ngo
L
et al
;
SAGES Study Group
:
Association of plasma neurofilament light with postoperative delirium.
Ann Neurol
2020
;
88
:
984
94
18.
Khalil
M
,
Teunissen
CE
,
Otto
M
et al
:
Neurofilaments as biomarkers in neurological disorders.
Nat Rev Neurol
2018
;
14
:
577
89
19.
Leung
JM
,
Rojas
JC
,
Tang
C
et al
;
and the Perioperative Medicine Research Group
:
Presence of preoperative neurodegeneration biofluid markers in patients with postoperative delirium.
Anesthesiology
2023
;
139
:
432
43
20.
Reinsfelt
B
,
Ricksten
SE
,
Zetterberg
H
,
Blennow
K
,
Fredén-Lindqvist
J
,
Westerlind
A
:
Cerebrospinal fluid markers of brain injury, inflammation, and blood-brain barrier dysfunction in cardiac surgery.
Ann Thorac Surg
2012
;
94
:
549
55
21.
DiMeglio
M
,
Furey
W
,
Hajj
J
et al
:
Observational study of long-term persistent elevation of neurodegeneration markers after cardiac surgery.
Sci Rep
2019
;
9
:
7177
22.
Saller
T
,
Petzold
A
,
Zetterberg
H
et al
:
A case series on the value of tau and neurofilament protein levels to predict and detect delirium in cardiac surgery patients.
Biomed Pap Med Fac Univ Palacky Olomouc Czech Repub
2019
;
163
:
241
6
23.
Alifier
M
,
Olsson
B
,
Andreasson
U
et al
:
Cardiac surgery is associated with biomarker evidence of neuronal damage.
J Alzheimers Dis
2020
;
74
:
1211
20
24.
Barbu
M
,
Jónsson
K
,
Zetterberg
H
et al
:
Serum biomarkers of brain injury after uncomplicated cardiac surgery: Secondary analysis from a randomized trial.
Acta Anaesthesiol Scand
2022
;
66
:
447
53
25.
Wiberg
S
,
Holmgaard
F
,
Zetterberg
H
et al
:
Biomarkers of cerebral injury for prediction of postoperative cognitive dysfunction in patients undergoing cardiac surgery.
J Cardiothorac Vasc Anesth
2022
;
36
:
125
32
26.
Khalifa
C
,
Lenoir
C
,
Robert
A
et al
:
Intra-operative electroencephalogram frontal alpha-band spectral analysis and postoperative delirium in cardiac surgery: A prospective cohort study.
Eur J Anaesthesiol
2023
;
40
:
777
87
27.
McDonagh
DL
,
Mathew
JP
,
White
WD
et al
;
Neurologic Outcome Research Group
:
Cognitive function after major noncardiac surgery, apolipoprotein E4 genotype, and biomarkers of brain injury.
Anesthesiology
2010
;
112
:
852
9
28.
Kuhle
J
,
Barro
C
,
Andreasson
U
et al
:
Comparison of three analytical platforms for quantification of the neurofilament light chain in blood samples: ELISA, electrochemiluminescence immunoassays and Simoa.
Clin Chem Lab Med
2016
;
54
:
1655
61
29.
Andreasson
U
,
Perret-Liaudet
A
,
van Waalwijk van Doorn
LJ
et al
:
A practical guide to immunoassay method validation.
Front Neurol
2015
;
6
:
179
30.
Chanques
G
,
Garnier
O
,
Carr
J
et al
:
The CAM-ICU has now a French “official” version. The translation process of the 2014 updated Complete Training Manual of the Confusion Assessment Method for the Intensive Care Unit in French (CAM-ICU.fr).
Anaesth Crit Care Pain Med
2017
;
36
:
297
300
31.
Chanques
G
,
Jaber
S
,
Barbotte
E
et al
:
Validation of the French translated Richmond vigilance-agitation scale.
Ann Fr Anesth Reanim
2006
;
25
:
696
701
32.
Inouye
SK
,
van Dycke
C
,
Alessi
C
,
Balkin
S
,
Siegal
A
,
Horwitz
R
:
Clarifying confusion: The Confusion Assessment Method.
Ann Int Med
1990
;
113
:
941
8
33.
Inouye
SK
,
Leo-Summers
L
,
Zhang
Y
,
Bogardus
ST
, Jr
,
Leslie
DL
,
Agostini
JV
:
A chart-based method for identification of delirium: Validation compared with interviewer ratings using the Confusion Assessment Method.
J Am Geriatr Soc
2005
;
53
:
312
8
34.
Akamine
S
,
Marutani
N
,
Kanayama
D
et al
:
Renal function is associated with blood neurofilament light chain level in older adults.
Sci Rep
2020
;
10
:
20350
35.
Koini
M
,
Pirpamer
L
,
Hofer
E
et al
:
Factors influencing serum neurofilament light chain levels in normal aging.
Aging (Albany NY)
2021
;
13
:
25729
38
36.
Desai
P
,
Dhana
K
,
DeCarli
C
et al
:
Examination of neurofilament light chain serum concentrations, physical activity and cognitive decline in older adults.
JAMA Netw Open
2022
;
5
:
e223596
37.
Celikbilek
A
,
Tanik
N
,
Sabah
S
et al
:
Elevated neurofilament light chain (NFL) mRNA levels in prediabetic peripheral neuropathy.
Mol Biol Rep
2014
;
41
:
4017
22
38.
Hendy
A
,
Hall
R
:
Cardiac surgery and the blood-brain barrier.
Anesthesiol Clin
2019
;
37
:
787
800
39.
Subramaniyan
S
,
Terrando
N
:
Neuroinflammation and perioperative neurocognitive disorders.
Anesth Analg
2019
;
128
:
781
8
40.
Kalm
M
,
Bolström
M
,
Sandelius
A
et al
:
Serum concentrations of the axonal injury marker neurofilament light protein are not influenced by blood-brain barrier permeability.
Brain Res
2017
;
1668
:
12
9
41.
Shann
KG
,
Likosky
DS
,
Murkin
JM
et al
:
An evidence-based review of the practice of cardiopulmonary bypass in adults: A focus on neurologic injury, glycemic control, hemodilution, and the inflammatory response.
J Thorac Cardiovasc Surg
2006
;
132
:
283
90
42.
Danielson
M
,
Wiklund
A
,
Granath
F
et al
:
Association between cerebrospinal fluid biomarkers of neuronal injury or amyloidosis and cognitive decline after major surgery.
Br J Anaesth
2021
;
126
:
467
76
43.
Taylor
J
,
Eisenmenger
L
,
Lindroth
H
et al
:
Perioperative ischaemic brain injury and plasma neurofilament light: A secondary analysis of two prospective cohort studies.
Br J Anaesth
2023
;
130
:
e361
9
44.
Brown
CH
,
Lewis
A
,
Probert
J
et al
:
Perioperative neurofilament light plasma concentrations and cognition before and after cardiac surgery: A prospective nested cohort study.
Anesthesiology
2022
;
137
:
303
14
45.
Khalil
M
,
Pirpamer
L
,
Hofer
E
et al
:
Serum neurofilament light levels in normal aging and their association with morphologic brain changes.
Nat Commun
2020
;
11
:
812
46.
Abelha
D
,
Fernandes
V
,
Botelho
M
et al
:
Apolipoprotein E e4 allele does not increase the risk of early post-operative delirium after major surgery.
J Anesth
2012
;
26
:
412
21
47.
Vasunilashorn
SM
,
Ngo
LH
,
Inouye
SK
et al
:
Apolipoprotein E genotype and the association between C-reactive protein and postoperative delirium: Importance of gene-protein interactions.
Alzheimers Dement
2020
;
16
:
572
80
48.
Dworkin
A
,
Lee
DS
,
An
AR
,
Goodlin
SJ
:
A simple tool to predict development of delirium after elective surgery.
J Am Geriatr Soc
2016
;
64
:
e149
53
49.
Gailiušas
M
,
Andrejaitiene
J
,
Širvinskas
E
,
Krasauskas
D
,
Švagždiene
M
,
Kumpaitiene
B
:
Association between serum biomarkers and postoperative delirium after cardiac surgery.
Acta Med Litu
2019
;
26
:
8
10
50.
Ballweg
T
,
White
M
,
Parker
M
et al
:
Association between plasma tau and postoperative delirium incidence and severity: A prospective observational study.
Br J Anaesth
2021
;
126
:
458
66