2. Risk Factors
Randomized results have shown that the incidence of POD is related to various perioperative risk factors. Knowledge of risk factors can help in clinical decision making and the identification of high-risk patients. The risk factors correlated with the onset of delirium, such as type of anesthesia, age, preoperative cognitive function, neurological function, and environmental factors, are summarized.
2.1. Anesthesia
The effect of the different types of anesthesia on POD is controversial. In a previous study, regional anesthesia and analgesia did not show any benefit with respect to POD compared with general anesthesia
[12]. Recently, emerging evidence has been published to provide additional information on this topic. Regional anesthesia, such as caudal block, fascia iliac compartment block, and intravertebral anesthesia, is also mentioned to reduce the incidence of POD
[13][14][15]. A recent study used topological data analysis to assess the phenotypic subgroups of delirium and suggested that elderly patients are more susceptible to POD and that this influence may be amplified by regional anesthesia
[16]. In a recent study, data showed that regional anesthesia significantly reduced POD incidence and severity. The positive effect of regional anesthesia was especially reflected in pediatric patients rather than elderly patients on postoperative days 1–5
[17]. In neurosurgical operations, a longer duration of anesthesia is expected to induce POD due to the impairment of neurons. Patients with anesthesia for more than 4 h showed a dramatically higher incidence of POD. Interestingly, a history of anesthesia did not affect the occurrence of POD
[18].
2.2. Age
POD is the most common complication following surgery in elderly patients. The incidence of POD has been estimated to range from 4% to 53% following fracture surgery in the elderly
[19]. It is hypothesized that the increased risk of imbalance in cortical neurotransmitters or inflammatory responses contributes to delirium. In spinal surgery, patients younger than 73 years had a significantly lower incidence of delirium. Older age, low preoperative cognitive function, longer duration of surgery, and transfusion are important risk factors for POD
[20]. Similarly, older age is also a critical intrinsic predictor of POD in those patients undergoing transcranial surgery. One study included patients aged 14–80 years who underwent brain tumor resection. The mean age of patients without and with POD was 47.1 ± 14.28 and 51.9 ± 12.66, respectively
[18].
2.3. Cognitive Condition
Patients who undergo surgery and anesthesia are also at high risk of temporary or permanent cognitive impairment due to acute stress. Mounting evidence suggests that delirium may lead to permanent cognitive decline and dementia in some patients. Cognitive impairment was common (>50%) in surgical patients who developed POD, and the impairment persisted up to 1-year postoperatively
[21]. Consistently, the cognitive state before surgery is a strong predictor of POD occurrence. Previous cognitive dysfunction, such as dementia or neurodegenerative status, greatly increases the vulnerability to POD
[22][23]. In elderly patients with dementia, delirium is associated with increased rates of cognitive decline, admission to institutions, and mortality
[21].
2.4. Intrinsic and Extrinsic Factors
The development of POD is multifactorial. Despite its main etiological factors, multiple intrinsic and extrinsic elements also promote the progression of POD. POD in neurosurgery patients can be induced due to hypothalamic syndrome, infection, electrolyte disturbances, fever, and profuse urination. A longer duration of ICU patients’ stay in the ICU was also associated with a higher incidence of POD
[18]. In particular, cardiac arrest is an additional predisposing factor for POD because most survivors of cardiac arrest are treated in the ICU. A longer stay in the ICU has been hypothesized to induce POD. The occurrence of delirium prolongs the duration of the ICU and hospital stay and adversely affects functional outcomes
[24].
Surgeons and anesthesiologists are devoting great effort to the prediction of the occurrence and the minimization of the risk of POD. The monitoring of processed electroencephalograms (EEGs) is considered an alternative method to predict the occurrence of POD. One moderate-quality evidence study indicated that EEG-optimized anesthesia could reduce the risk of POD in those patients aged ≥ 60 years who underwent non-cardiac or non-neurosurgical procedures
[25]. However, a more advanced study did not agree with this conclusion. They found that in older adults who underwent major surgery, EEG-guided anesthetic administration, compared with usual care, did not decrease the incidence of POD
[26].
3. Pathological Theories
Taking recent insights into comprehensive consideration, researchers conclude with five dominant pathological theories, as illustrated in Figure 1A, which may explain the occurrence and development of POD characterized by disturbances in attention, awareness, and cognition.
Figure 1. Schematic diagram of pathological mechanisms underlying postoperative delirium: (A) five dominant pathological theories that may account for the occurrence and development of POD characterized by disturbances in attention, awareness, and cognition; (B) S100A8, as a main member of DAMPs, promotes the activation of TLR4 in macrophages and microglia and then increases the expression of TNF-α; TNF-α will bind to TNFR on endothelial cells, subsequently triggering their necroptosis, which disrupts BBB’s integrity and increases BBB’s permeability.
3.1. Neuroinflammation
Inflammation is inevitable after surgery as a protective response to injury. However, peripheral inflammation may trigger neuroinflammation, leading to the dysfunction of the central nervous system (CNS) and the subsequent neurobehavioral and cognitive symptoms of postoperative delirium. Additionally, inflammation from the periphery to the CNS starts with increased permeability of the blood–brain barrier (BBB).
The BBB is a highly regulated and maintained interface that separates the peripheral circulation from the CNS. A specific monolayer of endothelial cells, which forms the capillaries of the brain, is the main component of the BBB. Other components of the BBB anatomy include astrocytes, pericytes, neurons, and extracellular matrix
[27][28].
Cellular injury caused by aseptic surgical trauma can induce the release of damage-associated molecular patterns (DAMPs) that activate the peripheral innate immune system
[29]. S100A8 (Migration suppressor-associated protein-8 (MRP8)) is an important proinflammatory cytokine in many inflammatory conditions and is expressed in large quantities by activated neutrophils and monocytes
[30][31]. As a main member of the DAMPs, S100A8 has been shown to promote the activation of Toll-like receptor 4 (TLR4) in macrophages and microglia
[32][33][34]. In the TLR4 signaling pathway, MyD88 is an important activator of the NF-κB signaling pathway
[35][36]. Previous studies have confirmed that S100A8 induced by surgery activates the TLR4/MyD88 pathway in mouse models
[33]. A recent study in a rat model showed that S100A8/A9 binds to TLR4 and increases the expression of matrix metalloproteinases (MMPs), tumor necrosis factor α (TNF-α), and IL-6 (interleukin-6) through the NF-κB signaling pathways in nucleus pulposus cells, which contributes to inflammation-related pain
[37].
The upregulation of TNF-α has been widely detected in postoperative patients
[38][39]. TNF-α secreted by activated microglia cells binds to the TNF receptor (TNFR) on endothelial cells, subsequently triggering necroptosis through receptor-interacting protein kinase 1 (RIPK1), RIPK3, and mixed-lineage kinase domain-like pseudo kinase (MLKL), which disrupts the integrity of BBB and increases the permeability of BBB
[40][41], as illustrated in
Figure 1B. Moreover, TNF-α induces the release of MMP-9 from pericytes, resulting in increased endothelial permeability in BBB models in vitro
[42].
The proinflammatory cytokine IL-6 is significantly upregulated in patients after orthopedic surgery
[38][39][43]. In mouse models, IL-6 disrupts the BBB and promotes hippocampal inflammation through bone marrow-derived monocytes
[44].
Upon infiltration into the brain parenchyma through the BBB, peripheral factors such as DAMPs and cytokines can trigger downstream neuroinflammatory responses.
Microglial cells, which account for 20% of the total glial cell population of the brain, are the main components of the posterior gray and white matter
[45] and function to monitor the well-being of their environment and maintain homeostasis through innate defense mechanisms or specific immune reactions
[46]. In a normal CNS environment, microglia are shut down with a scarce expression of many typical proteins on the surface of other tissue macrophages
[47][48]. The activation states of microglial cells can be divided into M1 (classic activation) triggered by interferon-γ and M2 (alternative activation), which are mainly triggered by the Th2 cytokines IL-4 and IL-13
[49][50][51]. Activated microglial cells have long been considered the main source of proinflammatory factors such as cytokines, eicosanoids, complement factors, excitatory amino acids, reactive oxygen radicals, and nitric oxide
[46][52]. Furthermore, owing to the microglial-secreted cytokines, the neurotoxic reactive subtype of astrocytes, astrocytes, can be induced by classically activated microglial cells, which contribute to the death of neurons and oligodendrocytes
[53].
A recent study showed that neuronal dysfunction after traumatic brain injury, including disrupted neuronal homeostasis, reduced dendritic complexity, and defective compound action potential, can be attenuated by microglial elimination
[54]. Moreover, in those patients undergoing abdominal surgery, microglial activation has been detected and associated with impairments in cognitive function
[54]. Meanwhile, the endogenous mesencephalic astrocyte-derived neurotrophic factor has been shown to have positive effects on POD by inhibiting surgery-induced inflammation and microglial activation
[55][56].
3.2. Oxidative Stress
Oxidative stress is an imbalance between the production of oxidants in cells and tissues and the particular biological processes that trigger the detoxification of these reactive products
[57]. Under physiological and pathological conditions, cells can produce free radicals and reactive oxygen species (ROS) through the NADPH–oxidase system, xanthine oxidase, and mitochondrial electron transport chain
[58].
The disruption of the BBB is a common cause of oxidative stress. Rat brain microvascular endothelial cells exposed to hydrogen peroxide (H
2O
2) at high concentrations display significant monolayer hyperpermeability with decreased cell viability and induced apoptosis
[59]. It has been found that ROS can result in the BBB’s cytoskeleton rearrangement and redistribution and disappearance of tight junction (TJ) proteins, mediated by RhoA, PI3 kinase, and PKB signaling
[60]. Occludin, a component of intercellular TJ protein complexes, moves away from TJ during oxidative stress
[61]. Furthermore, toxic cell H
2O
2 concentrations cause occludin cleavage with the involvement of MMP-2
[62], and the upregulation of MMP-9 induces posttraumatic nerve and BBB injury, which may be partially mediated by Scube2 and SHH through the hedgehog pathway
[63]. In turn, patients with increased disruption of the BBB are more vulnerable to neuronal injury induced by oxidative stress
[64].
Oxidants have been found to increase the gating potential of the mitochondrial permeability transition pore (mPTP), resulting in hypersensitivity to Ca
2+ activation during neuronal oxidative stress
[65]. The high permeability of the mitochondrial membrane induced by mPTP dysfunction not only impairs the mitochondrial electron transport chain but also causes mitochondrial swelling, consequently leading to the over-release of ROS and neuronal necrosis
[66]. Cyclosporine A, an inhibitor of mPTP opening that suppresses oxidative stress
[67][68], was found to attenuate delirium-like behavior induced by anesthesia and surgery, with decreased ROS levels in the hippocampus of POD-like mice
[69]. Meanwhile, as an innate protective mechanism to combat invading pathogens, macrophage cell lines, including microglia, can produce superoxide radicals and nitric oxide, resulting in the production and spread of peroxynitrite
[70]. Peroxynitrite has been shown to induce neuronal apoptosis through the intracellular release of zinc and subsequent activation of p38 mitogen-activated protein kinase and caspase 3
[71].