Sepsis is defined as a life-threatening organic dysfunction caused by a dysregulated host response to infection [1]. It was estimated that in 2017, there were 48.9 million cases and 11 million sepsis-related deaths worldwide, with a mortality rate that varies from 26% to 60% depending on the severity [2]. Despite all the efforts to improve sepsis treatment in the last decades, there is no specific therapy for sepsis, and mortality remains high, especially in low- and middle-income countries.

Sepsis is a clinical syndrome characterized by a maladaptive host response to infection that may be significantly amplified by endogenous factors. The pathogenesis of sepsis goes beyond an uncontrolled inflammatory response and includes major modifications in nonimmunologic systems such as cardiovascular, autonomic, neuroendocrine, coagulation, bioenergetic, and metabolic alterations [1]. While the inflammatory response in sepsis is relatively well-understood, the mechanisms driving metabolic deregulation and multiple organ dysfunction remain puzzling.

Multiple organ dysfunction is a hallmark of sepsis pathogenesis, and neurological manifestations are a frequent and underestimated symptom. Sepsis induces an acute brain dysfunction that is not related to direct brain infection and is characterized by various clinical and electroencephalographic changes ^[3][4][3,4]. This acute brain dysfunction is known as sepsis-associated encephalopathy (SAE), and it ranges from sickness behavior to altered consciousness, varying from confusion to delirium and, in worse cases, even coma [5]. SAE is strongly associated with higher mortality [6] and long-term cognitive impairments ^[7][8][7,8] affecting patients’ quality of life. To date, no curative or preventive treatment for SAE exists, and some pharmacological attempts have proved ineffective or even dangerous [9]. The absence of an effective therapeutic strategy is explained by the lack of precise knowledge of pathophysiological mechanisms. Some authors even suggest that the neurocognitive sequelae identified in the patients who had stayed in an ICU could be a risk factor for progressive neurodegenerative disorders of higher functions ^[10][11][10,11].

The mechanisms involved in SAE are diverse and include neurotransmitter dysfunction, inflammatory and ischemic lesions to the brain, microglia activation, changes in the blood–brain barrier (BBB) permeability, oxidative stress due to inflammation and mitochondrial dysfunction, and bioenergetic shifts that are part of metabolic adaptations to systemic inflammation ^[12][13][12,13]. Evidence suggests that alterations of the BBB resulting from both microglial and endothelial cell activation [14] could, in part, explain the acute neurological dysfunction by increasing the entry of inflammatory mediators and neurotoxic substances within the CNS [15]. Local production of pro-inflammatory markers and damage-associated molecular patterns (DAMPs) by the cells maintaining BBB (activated endothelial cells and astrocytes and perivascular macrophages) also contribute to neuroinflammation. Dysregulated cytokine responses are significant contributors to tissue injury and neurological deficits ^[16][17][18][16,17,18]. Inflammatory cytokines are involved in the pathophysiology of intensive care unit-acquired weakness, mostly, of neuropathies and myopathies [19]. All these mechanisms constitute a process known as neuroinflammation, which is the focus of this review.

To some extent, neuroinflammation is essential for maintaining brain homeostasis by inducing repair mechanisms. However, non-resolving neuroinflammation is a potentially harmful mechanism of brain damage. In this respect, microglia are major players in neuroinflammation, and exacerbated microglia activation has been associated with neurodegenerative diseases. As neurodegenerative diseases progress, microglia switch from a helpful role to a dysfunctional phenotype that becomes detrimental to neurons [20]. It is becoming clear that microglia surrounding Aβ plaques or Lewy bodies, for example, are not activated, but are nonfunctional [21]. There is a delicate balance determining whether microglia—and neuroinflammation in general—have beneficial or detrimental effects on the brain. This dual role is covered in the following sections, with a focus on sepsis and SAE.