Sepsis and Systemic Inflammatory Syndromes

Key points

•
Sepsis-associated encephalopathy (SAE) is a commonly occurring syndrome and major risk factor for long-term cognitive impairment and mental health problems.
•
The underlying biology of SAE is multifactorial, comprising systemic-brain immune crosstalk, blood-brain barrier disruption, glial activation, and synaptic injury.
•
Neuroinflammation, a key immunologic process in SAE, involves variable region-specific activation of diverse immune cells, making accurate quantification difficult.

Abbreviations

ACE	arachnoid cuff exit
BBB	blood-brain barrier
BCSFB	blood–cerebrospinal fluid barrier
CNS	central nervous system
CSF	cerebrospinal fluid
DCE-MRI	Dynamic contrast-enhanced MRI
EEG	electroencephalography
GCS	Glasow Coma Scale
IFNs	Interferons
IL	Interleukin
NFL	neurofilament light chain
PAMP	pathogen-associated molecular patterns
PRR	pattern recognition receptors
PSS	postsepsis syndrome
PTSD	posttraumatic stress disorder
ROS	reactive oxygen species
SAE	sepsis associated encephalopathy
SV2	synaptic vesicle glycoprotein 2
TNF-α	tumor necrosis factor alpha
TSPO	translocator protein

Introduction

Sepsis is defined as a life-threatening organ dysfunction due to a dysregulated host response to infection and is characterized by immune dysregulation ranging from hyperinflammation to profound immunosuppression. Sepsis affects one in 3 patients in the intensive care unit (ICU) and is one of the leading causes of death worldwide with high mortality rates; 27% in hospitalized patients and up to 42% in ICU patients. Among survivors, the 5-year survival varies between 44% and 62%. ^, Sepsis often causes multiorgan failure, including the kidneys, lungs, and liver. The brain is also commonly affected, with dysfunction occurring both in the acute phase as well as in the long-term postsepsis phase. The incidence of clinical characteristics of brain dysfunction and its biological underpinnings related to the immune-brain interplay will be described.

The clinical syndrome in the acute and long-term phase

Sepsis-associated encephalopathy (SAE) is a neurologic syndrome characterized by (diffuse) dysfunction of the brain caused by sepsis ( Fig. 1 ). In the acute phase, patients with SAE typically present with an acute onset of encephalopathy ranging from delirium to altered mental states and even coma. ^, Other less frequently encountered symptoms are agitation (10%), dysautonomia, convulsive seizures (2%), and focal neurologic signs (1%). Other risk factors such as hyperglycaemia and hypoglycaemia, hypercapnia, hypernatremia, central nervous system (CNS) infection, bacteraemia, may account for these uncommon presentations, rather than being caused by sepsis per se. Without established diagnostic biomarkers or tools, the diagnosis remains purely clinical after excluding other possible causes. This makes the diagnosis of SAE less accurate and at risk to bias, as shown by the wide range of reported SAE incidence, which generally ranges between 30% and 70% of patients with sepsis. ^,

Alt text begins. A schematic diagram shows systemic inflammation progressing to neuroinflammation, neuro-axonal injury, cognitive deficits, and mental health deterioration. Alt text provided by Elsevier. — Fig. 1

Sepsis-Associated Encephalopathy and Survival

In a study with over 2500 patients with sepsis at ICU admission, of whom approximately half had SAE, even mild alterations of mental status, defined by a Glasow Coma Scale (GCS) of 13 or 14, were associated with a 40% higher chance of dying. The degree of SAE severity was also correlated with a more than 3 times increased risk of dying, with patients having a GCS of 3 to 8 showing the worst prognosis. Furthermore, presence of SAE was independently associated with increased need and duration of mechanical ventilation, higher need for sedatives, increased organ supportive therapies such as renal replacement therapy, vasopressor use, and increased ICU length of stay. A recent meta-regression analysis with more than 20.000 patients with sepsis confirmed that patients with SAE at ICU admission had a higher disease severity and longer ICU length of stay.

Postsepsis Syndrome

The long-term sequelae of sepsis are defined as postsepsis syndrome (PSS), which consists of a combination of long-term deterioration of the physical state, mental health disorders, and cognitive decline. Compared to nonseptic patients, patients who survive sepsis develop a higher rate of new functional limitations, in patients who had preexisting limitations as well as patients who did not have any functional limitation beforehand. Importantly, these declines in physical function can persist for many years, impacting independence, mental wellbeing, and quality of life. ^,^, Later, we will focus in more detail on outcomes related to brain dysfunction.

Postsepsis Mental Health

Patients surviving sepsis frequently suffer from subsequent depression, anxiety, posttraumatic stress disorder (PTSD) and sleep disturbances and these sequelae are more likely following SAE. In a cohort of ICU survivors with sepsis or septic shock, anxiety, depression, and PTSD symptoms were present in 67%, 49%, and 46% of patients, respectively, shortly after ICU discharge. After 1 year, 38% of anxiety, 50% of depression, and 31% of PTSD symptoms remained. More severe depression postsepsis was associated with higher levels of PTSD and lower health-related quality of life. PTSD is particularly prevalent in sepsis survivors, with pooled prevalence of 20% within the first year after ICU discharge reported in a meta-analysis of 48 studies. Importantly, PTSD in ICU survivors has been associated with significantly worse functional outcomes, including higher rates of moderate-to-severe disability and more pronounced depressive and anxiety symptoms. In addition to depression and PTSD, survivors frequently report persistent anxiety, fear-related symptoms, and sleep disturbances, reflecting the psychological impact of severe illness and ICU admission. Collectively, these findings demonstrate that mental health disorders are prevalent and persistent after sepsis, and are strongly associated with impaired quality of life.

Postsepsis Cognition

The prevalence of moderate to severe cognitive impairment increased 11% postsepsis and up to a fifth of sepsis survivors shows long-term cognitive deficits. In contrast to patients with sepsis, nonsepsis general hospitalizations did not show a significant change in moderate to severe cognitive impairment. Patients with SAE have an even higher risk of long-term cognitive decline compared to sepsis survivors without SAE. ^,^, Within the group of SAE, delirium occurs in up to a third of patients with sepsis and has been shown to be the strongest predictor of long-term cognitive decline.

Pooled data from 8 studies and with follow-up periods up to 11 years, showed that sepsis was associated with a 62% increased risk to develop dementia. This risk was doubled in patients with severe sepsis. In patients with normal neuropsychological assessments, the occurrence of sepsis, compared to patients with no or less severe infectious episodes, was associated an 82% increased risk to develop dementia within the follow-up period. These studies collectively show that sepsis and severity of sepsis are important risk factors for the development of subsequent dementia.

Importantly, the risk of developing dementia postsepsis is strongly linked to higher age. Aging is accompanied by cumulative damage due to repetitive inflammatory hits and due to physiologic age-related decline in cognitive function. This leads to a low-grade, chronic systemically inflamed state, which has been coined as inflammaging . Within the brain, this consists of chronically low grade activated ( primed ) (micro)glial cells, accumulation of aged glial cells and sustained expression of proinflammatory cytokines. Apart from the decline in functional reserves, the inflammaging associated primed neuroimmune state may explain why people of higher age are more likely to develop persistent postsepsis brain dysfunction. Their primed neuroimmune cells mount a more severe and sustained neuroinflammatory response to a sepsis. This is corroborated by animal studies, which showed that primed microglia respond more vigorously to systemic inflammation. In addition, elderly are more likely to already have (undetected) dementia related protein aggregates in the brain, such as amyloid-β, α-synuclein or TAU protein, where additional neuroaxonal loss due to sepsis, accelerates the clinical diagnosis of dementia.

Although sepsis severity and aging are the most important risk factors for developing SAE and subsequent postsepsis cognitive impairments, additional predisposing factors such as (mild) preexisting cognitive deficits, previous acquired severe illnesses, comorbidities, preexistent depression, and metabolic disturbances also strongly add to this risk. ^,^,

Pathophysiology

The pathophysiology of SAE is complex with detailed underlying biological pathways remaining largely unknown. In sepsis, circulating immune cells (neutrophils and monocytes) as well as tissue macrophages are activated through binding of pathogen-associated molecular patterns (PAMPs) to pattern recognition receptors (PRRs), leading to release of proinflammatory cytokines such as tumor necrosis factor alpha (TNF-α), interleukin (IL)-1, IL-6, IL-8, IL-12, IL-18 and type I and II interferons (IFNs), including IFN-γ. These cytokines activate nearby and distal immune cells and this chain reaction may result in a rapid hyperinflammatory immune response. Over recent decades, the brain was shown to be more immunologically active than previously assumed with bidirectional communication between the brain and systemic immune compartments. In patients with sepsis, systemic (hyper) inflammation is hypothesized to dysregulate crosstalk between the systemic and brain compartment. The immune-brain crosstalk occurs at distinct anatomic sites in the brain, which include the blood-brain barrier (BBB), the outer border regions of the brain, and the blood–cerebrospinal fluid barrier (BCSFB). ^, Apart from being crosstalk sites, these regions were shown to be important parts of the brain’s immune response. Cerebrospinal fluid (CSF) plays an important role in the immune crosstalk at several brain border regions.

The Immune-Brain Crosstalk Regions

Blood–brain barrier

Under normal conditions, the BBB regulates the transfer of molecules between the circulatory system and the CNS and is very selectively permeable in part due to its layer of endothelial cells with tight junctions ( Fig. 2 ). During sepsis several processes can disrupt this barrier. First, cytokines damage the endothelium affecting its tight junctions, leading to microvascular leakages and cellular edema. Autopsy studies in patients who died of sepsis showed reduced tight junction proteins in BBB endothelial cells. Second, BBB endothelial cells enhance their expression of intracellular adhesion molecules, increasing leukocyte adhesion to endothelium and promoting leukocyte migration toward the brain. Third, radicals such as reactive nitrogen species and reactive oxygen species (ROS), and matrix metalloproteinases produced during sepsis further amplify the damage to the BBB. This leads to activation of astrocytes and especially microglia within the brain parenchyma, further contributing to BBB breakdown through disruption of the glia limitans and through facilitating transport of immune cells in the neuropil. In the initial stages of systemic inflammation, these processes occur on the luminal side of the BBB, but once the BBB permeability is compromised and inflammatory mediators penetrate into the brain, the BBB may be further disrupted from the abluminal side.

Outer brain border regions

Over the last decade, brain border regions, including the skull bone marrow and meninges, emerged as sites of active crosstalk between the peripheral and brain’s immune systems. Beneath the skull bone, there are 3 meningeal layers that separate the skull from the brain parenchyma ( Fig. 3 ). These layers allow for bidirectional communication between the peripheral immune system, the brain immune system, and the brain parenchyma. CSF is a key transporter of this communication, as it contains signaling proteins, waste products and cells that originate from the brain parenchyma or were in close contact with it.

Alt text begins. A meningeal flow diagram shows cerebrospinal fluid movement, arachnoid cuff exit points, immune cells, bone marrow, and dural sinus drainage. Alt text provided by Elsevier. — Fig. 3

Blood vessels of the brain have a perivascular compartment that contains CSF that is in contact with the brain parenchymal interstitium. CSF flow through the interstitium provides nutrients to brain cells, and drains signaling and waste products from the interstitial space of the brain parenchyma (see Fig. 3 ). The mechanism of CSF-interstitial fluid exchange is now regarded as an important part of the glymphatic system of the brain. The CSF flow through the brain parenchyma is regulated by parenchymal border macrophages, also modulating movement of cytokines and antigens. Disruption of this glymphatic flow may increase neuroinflammation. Systemic inflammation was shown to disrupt this glymphatic system. ^, Impaired CNS waste removal, indirectly influenced by parenchymal border macrophages, has also been associated with Alzheimer’s disease. ^, In SAE, increased accumulation of Alzheimer’s related protein aggregates (amyloid-β and phosphorylated TAU) was found in both animal and human studies. ^,

The brain’s blood vessels travel through the brain parenchyma and cross the meninges where some veins end up crossing into the skull (see Fig. 3 ). The bridging veins can relay signals from CSF in their perivascular space via the dura mater to the skull bone marrow through arachnoid cuff exit (ACE) points. ^, Additionally, immune cells travel from bone marrow to meninges by channels in the bone marrow and meninges. Thus, bidirectional communication between bone marrow and blood in the skull and the brain is an important component of neuroinflammation, affecting brain immunology and neuronal functioning (see Fig. 3 ). After experimental murine sepsis, skull bone marrow showed decreased amounts of neutrophils, while neutrophil efflux from skull bone marrow into meninges was increased, which enhanced neuroinflammation.

Cerebral veins drain in dural sinuses, which are located in the dura mater of the meninges. In the vicinity of these sinuses, lymphatic vessels of the brain are located. These sites have been shown to be immunologically active hotspots that include various immune cells. Meningeal lymphatics contribute to the drainage of CSF and CNS waste products, and allow immune cells to enter draining lymph nodes. In the meninges, lymphocytes and antigen-presenting cells can present antigens of brain parenchyma encountered in the CSF to T-cells, and can thereby contribute to neuroinflammation. Recently, meninges were found to aid in the removal of CNS-reactive immature B-cells and to be a place for further development of immature B-cells which drifted through channels from the skull bone marrow. ^, During neuroinflammation, these immature B-cells were able to mature locally in the meninges, contributing to neuroinflammation.

Blood–cerebrospinal fluid barrier

The choroid plexus, located within the ventricles of the brain, produces CSF and is the site of the BCSFB. The BCSFB consists of vascularized stroma surrounded by CSF producing epithelial cells that are connected by tight junctions ( Fig. 4 ). The choroid plexus, including its stromal blood vessels, lacks a tight barrier like the BBB. CSF is separated from the blood compartment by a fenestrated endothelial layer, loosely connective choroid plexus stroma and an epithelial layer of CSF producing cells. This facilitates immune-brain communication between both compartments (see Fig. 4 ). Besides supporting tissue, the choroid plexus stroma is immunologically active as it contains lymphoid and myeloid populations, including T-cells and macrophages. In murine sepsis models, systemically activated antigen presenting cells have been shown to infiltrate the choroid plexus and release IL-1β, triggering stromal cells to produce proinflammatory factors. CD4 ⁺ T-helper cells produce IFN-γ and promote transepithelial migration of preactivated T- and B-cells into the CSF. Systemic inflammation has been shown to compromise the tight junctions in the epithelial cells of the choroid plexus. Migration of immune cells and compromised tight junctions could thus promote neuroinflammation at this brain site during sepsis. In addition, oxidative stress was found to be a contributor not only to BBB disruption but also BCSFB disruption during sepsis.

Main Immune Cells in Immune–Brain Crosstalk in Sepsis

The glial cells, microglia and astrocytes, are the main resident tissue immune cells of the brain parenchyma, and maintain cerebral homeostasis and synaptic functioning. Disruption of these glial cells may cause synaptic dysfunction and synaptic apoptosis, leading to function loss. ^, The glial cells, and especially microglia, play an important role in neuroinflammatory processes and are presumed to be one of the key drivers of brain dysfunction during sepsis. This is supported by autopsy findings in patients with sepsis showing microglial activation in absence of bacterial translocation into the brain parenchyma. ^, Activated microglia amplify the proinflammatory response, damaging cellular membranes, and contribute to mitochondrial dysfunction. This leads to impaired neuronal functioning and neuronal death. ^, The NLR family pyrin domain containing 3 (NLRP3) inflammasome pathway in microglia is gaining much attention in neuroinflammatory research with several drug targets being studied currently in animal and human studies.

Another proposed mechanism by which microglia may cause synaptic dysfunction is through synaptic pruning, which is largely mediated through complement proteins C1q and C3 which give phagocytotic signals to microglia via the C3-C3aR axis. In normal brain development, microglia prune synapses to allow plasticity of the brain and eliminate unnecessary synapses. However, patients with sepsis exhibit extensive phagocytic active CD68 ⁺ microglia, ^,^, and C1q complement activation, suggesting enhanced synaptic pruning. In neurodegenerative diseases, enhanced synaptic pruning is thought to contribute to cognitive decline. It is likely that enhanced synaptic pruning in neuroinflammation could negatively impact cognitive function. This is supported by the finding that synaptic proteins in the CSF were downregulated in patients with infection-associated delirium compared to patients with an infection without delirium. Similar synaptic proteins were also found to be downregulated in the CSF of patients with Alzheimer’s disease. To summarize, activated microglia contribute to a proinflammatory milieu, BBB disruption, synaptic dysfunction and apoptosis, leading to transient or permanent brain dysfunction.

While astrocytes are understudied in sepsis, they also promote neuroinflammation by excretion of proinflammatory cytokines and by activating microglia, which both indirectly contribute to neuronal damage. Recently, early astrocyte reactivity was highlighted as a driver of neuroinflammation in sepsis. Activated astrocytes in a murine infection model have also been proposed to promote phagocytosis of synapses through the C3-C3aR axis by releasing C3. Due to this, and importantly, their role as one of the 3 layers that make up the BBB, more research is warranted into their role in neuroinflammation, maintenance of the integrity of the BBB, and neuronal and synaptic homeostasis.

A lot of neuroinflammation related research has focused on the contribution of glial cells, and especially microglia. In part, because much of this research was related to neurodegenerative diseases, where microglia are found near the dementia-related protein aggregates in damaged areas. However, recently it was shown that other immune cells, and especially T-cells, also play a key role via bidirectional communication between the systemic and brain immune compartments in the brain border regions. In addition, single-cell transcriptomics of CSF has shown that approximately 70% of the immune cells in CSF are T-cells, with the remaining being dendritic cells and distinct phenotypes of monocytes.

Measures for neuroinflammation

Measuring neuroinflammation is difficult, as it is not a single measurable entity. It is inferred from proxies reflecting glial activation (microglia, astrocytes), inflammatory mediators, blood-brain barrier dysfunction, immune cell trafficking and proliferation, and downstream neuroaxonal injury. Currently, no measure captures all these proxies simultaneously, but rather most measures correlate with these proxies. This is particularly relevant in postacute infection syndromes, such as in postsepsis patients, where a combination of all these mechanisms can underly the pathogenesis. ^,

The most direct way to define and measure neuroinflammation would be through well characterized blood and CSF biomarkers. Proinflammatory and anti-inflammatory cytokines such as IL-1β, IL-6, IL-8, TNF, and IL-10 as well as chemokines such as CCL2 (MCP-1) and CXCL10, reflect (neuro) inflammatory immune signaling. ^, Markers of neuroinflammation are elevated in the acute phase but can also be chronically dysregulated in patients with persistent cognitive complaints. CSF concentrations of the proinflammatory cytokines IL-6 and IL-8 are upregulated in patients with SAE. Additionally, IL-6 and IL-8 concentrations, but also IL-1β, IL-10, MCP-1, and TNF have been associated with development of delirium. ^,^, Even at baseline, elevated cytokine concentrations have been shown to hold predictive value for the development of delirium. ^, In patients with persistent cognitive complaints ( brain fog ), persistent elevations of IL-6 and CXCL10 have been observed. However, in systemic inflammation higher cytokine levels in the CSF will be seen irrespective of clinical brain involvement, due to influx of systemic inflammatory mediators into the brain compartment. ^,^,

Glial proteins, including sTREM2, YKL-40, and to a lesser extent also GFAP and S100b, reflect microglia and astrocyte activation, and thus could be used as more specific biomarkers of neuroinflammation. ^,^,^, Also, neuroaxonal injury markers can be interpreted as downstream consequences and injury due to neuroinflammation. These include neurofilament light chain (NFL), NSE, GFAP, S100 B, UCH-L1, and TAU (total or phosphorylated), all of which are associated with neuronal damage and injury. ^,^,^, They have been shown to be of diagnostic use in discriminating sepsis with and without associated encephalopathy. However, they do not specifically reflect ongoing neuroinflammation as neuroaxonal injury may also occur due to other mechanisms present in sepsis patients, for example, hypoxia or hypotension.

Accurate determination of neuroinflammation using biomarkers remains challenging. CSF sampling is of interest, as it plays a key role in the immune-brain interaction and is in direct anatomic and biochemical contact with CNS immune cells, the different immune brain border niches and brain parenchyma. The main method of obtaining CSF is through invasive lumbar punctures, limiting longitudinal sampling. Due to the relative inaccessibility of the brain, blood-based biomarkers reflecting the degree of neuroinflammation would be an important aid for clinical assessment and research. Unfortunately, blood concentrations correlate poorly to neuroinflammatory profiles. A fundamental limitation of blood-based biomarkers is the fact CSF-blood exchange of biomarkers is restricted and thus highly influenced by blood–brain barrier permeability in addition to peripheral metabolism, renal clearance, extracerebral sources, and anatomic distance to the CNS. As a result, the degree to which circulating biomarker levels reflect CNS inflammation varies substantially. ^, For neuroinflammatory processes, blood-based biomarkers have proven unreliable. This is slightly different for neuroaxonal or glial damage related biomarkers such as NFL and GFAP, which show moderate to strong correlations between CSF and plasma, while other biomarkers such as UCHL1 and TAU do not. For example, blood NFL concentrations reflect neuroaxonal damage and predict long-term cognitive dysfunction in patients with severe COVID-19. This highlights the usefulness of blood-based biomarkers related to neuroaxonal damage. In addition, imaging techniques can be used to demonstrate neuroinflammation. The translocator protein (TSPO), which is expressed on microglia, astrocytes, neurons, and peripheral macrophages, is the most frequently used target in imaging neuroinflammation. Expression of TSPO in the healthy brain is low, with expression increasing significantly in neuroinflammatory environments. The expression of TSPO can be imaged in vivo using PET-CT imaging with TSPO radioligand such as [ ¹¹ C]PK11195 (first generation) and [ ¹⁸ F]DPA-714, [ ¹¹ C]PBR28 (second generation). ^, PET TSPO imaging, however, is limited by a polymorphism in the TSPO gene (rs6971), which results in a conformational change that negatively impacts the binding capacity of many (radiolabeled) TSPO antagonists, making imaging a challenge in these patients. An additional important limitation of PET imaging is the lack of detailed spatial resolution and more importantly that it does not provide any insight into (detailed) biological pathways.

Promising new applications of PET-CT imaging include the quantification of synaptic density using radiotracers that bind to synaptic versicles, which are neurotransmitter-containing units located in the axon terminals. ^, These radiotracers bind to the synaptic vesicle glycoprotein 2 (SV2), which is highly expressed at presynaptic terminals, with 2 to 5 copies per synaptic vesicle, making it a suitable biomarker for synaptic density. ^, While early SV2 ligands showed low labeling yield resulting in a comparatively low resolution, ^, the more recent [ ¹⁸ F]-SDM-8 ligands showed high resolution synaptic density imaging. ^,^, Synaptic density could prove to be a promising marker for brain dysfunction in patients with sepsis as downregulation of synaptic proteins was previously shown in patients with SAE.

PET-CT imaging is often accompanied by MRI to identify brain and vascular wall lesions, white matter hyperintensities, atrophy patterns, vasogenic edema resulting from neuroinflammation, and BBB dysfunction. In patients with septic shock who developed acute brain dysfunction, MRI imaging has identified leukoencephalopathy and ischemic stroke. Dynamic contrast-enhanced MRI (DCE-MRI) detected BBB leakage associated with inflammatory states but was not able to distinguish between immune activation and vascular injury due to other sources.

Synaptic transmission is one of the first processes to fail during severe metabolic or hypoxic stress. Electroencephalography (EEG) shows cortical synaptic activity of neurons, and in patients with SAE, slowing of the background activity on EEG corresponded with the severity of SAE. As with other measures however, EEG does not measure neuroinflammation and it lacks detailed spatial resolution. It can be used as an additional screening or severity tool of brain dysfunction in combination with other measures.

At present, detailed characterization of neuroinflammation can be inferred from multimodal approaches combining CSF and blood-based and imaging biomarkers. These approaches individually lack specificity because no marker uniquely captures neuroinflammation. Temporal dissociation between systemic and brain immune responses limits blood-based readouts. Finally, even when biomarkers suggest systemic inflammation, CNS inflammation, (neuroaxonal) injury and dysfunction, a causal relationship between these processes cannot be assumed. In the future, multitracer imaging, combining for example, PET tracers for both synaptic density and microglia activation, could enable the assessment of multiple pathologies by in vivo serially capturing the trajectory of these processes in the different disease stages (acute and long-term). Additionally, large scale CSF and blood immune cell profiling using multiomic single-cell sequencing and large-scale proteomics can provide insight into the biological underpinnings of neuroinflammation, and relate these to subsequent injury, impaired mental and cognitive outcomes or even stratifying patients into biological categories at risk for developing long-term brain dysfunction.

Only gold members can continue reading. Log In or Register to continue

Anesthesia Key

Fastest Anesthesia & Intensive Care & Emergency Medicine Insight Engine

Sepsis and Systemic Inflammatory Syndromes

Key points

Introduction

The clinical syndrome in the acute and long-term phase