Advances in stroke-associated pneumonia in the era of revascularization therapy

Abstract

In recent decades, the advent of revascularization treatments (RVTs), such as intravenous thrombolysis (IVT) and endovascular therapy, has significantly improved clinical outcomes in patients with acute ischemic stroke (AIS). However, stroke-associated pneumonia (SAP) remains a common complication that has a negative impact on the prognosis. This review addresses the mechanisms underlying the development of SAP in patients with AIS, evaluates risk factors, and discusses therapeutic and preventive measures in the era of RVTs. Despite advances in acute stroke care, the incidence of SAP remains high and comparable between IVT and endovascular treatment (EVT), although EVT may confer specific procedural risks related to prolonged mechanical ventilation, anesthesia, and intensive care unit stays. Current prevention strategies, including screening for dysphagia, oral hygiene, and immunomodulation, show promise but require further validation. Future research should focus on integrating biomarkers and imaging markers for early prediction of SAP and developing targeted interventions to improve patient outcomes.

Introduction

Stroke affects millions globally each year, resulting in significant disability, and with population aging, the incidence and burden of stroke are expected to increase further.^[1–4] The advent of vascular reperfusion techniques, including intravenous thrombolysis (IVT) and endovascular treatment (EVT), and advances in immunotherapy have led to dramatic improvements in stroke management in recent decades and in the treatment of ischemic stroke.^[5–7] Thrombolytic agents, such as streptokinase and urokinase, discovered in the 1950s, and the introduction of recombinant tissue plasminogen activator (rt-PA) in the 1980s,^[8] led to significant advances in the treatment of the acute phase of ischemic stroke after publication of a study by the NINDS in 1995.^[9] The ECASS III trial, published in 2008, extended the treatment window for rt-PA to 4.5 h, while the development of mechanical thrombectomy devices, such as Solitaire and Trevo, occurred around 2010,^[10] and their usefulness was confirmed in pivotal trials, including MR CLEAN, ESCAPE, EXTEND-IA, REVASCAT, and SWIFT PRIME.^[11–15] The DAWN and DEFUSE-3 trials subsequently extended the treatment options to up to 24 h after stroke onset for selected patients.^[16,17] These advances have significantly improved patient outcomes and quality of life and transformed clinical practice. IVT and EVT are now well-established treatments for acute ischemic stroke (AIS) caused by large vessel occlusion (LVO)^[18,19] and have significant clinical benefits.^[20]

The focus in the era of revascularization treatment (RVT) has been primarily on clinical effectiveness, while comparatively little attention has been paid to the potential impact on other stroke-associated complications. Complications post-stroke, particularly infections, are common in the acute phase of stroke treatment. Studies indicate that post-stroke infections, particularly pneumonia, are associated with increased mortality and functional loss.^[21–23] Stroke-associated pneumonia (SAP) refers to a respiratory tract infection that occurs within the first week after a stroke^[24] and is associated with poor functional outcomes, longer hospital stays, and higher hospital costs.^[25–29] RVT was expected to reduce the risk of complications. However, recent studies indicate that EVT may pose additional procedural risks for pneumonia during hospitalization,^[30,31] largely attributable to prolonged mechanical ventilation, anesthesia-related effects, and intensive care unit stays.

This review addresses the mechanisms via which pneumonia develops in patients with AIS, identifies risk factors for its occurrence, and examines both current and novel therapeutic and preventive measures in the era of RVT. By analyzing these elements, the review aims to identify strategies that could mitigate adverse outcomes, minimize ongoing complications, and improve the long-term prognosis, thereby contributing to more effective paradigms for the treatment of stroke and pneumonia.

Definition

SAP was first reported by Hilker et al^[32] in 2003 using the US Centers for Disease Control diagnostic guidelines. The Pneumonia in Stroke Consensus group updated these criteria in 2014, defining SAP as a lower respiratory tract infection occurring within 7 days of stroke onset and distinguishing between probable SAP (without typical changes on a chest radiograph) and definite SAP (with clear radiographic changes).^[33] Despite these advances, diagnosis of SAP remains challenging because of the limited sensitivity of chest radiography, although high-resolution computed tomography of the chest has demonstrated improved accuracy in distinguishing SAP from other respiratory diseases.^[34]

Incidence

According to a comprehensive systematic review and meta-analysis, the incidence of SAP after AIS has remained constant over recent decades.^[35] Remarkably, despite major advances in acute stroke therapy and stroke care, there is no evidence of a decline in the incidence of SAP over time, probably because of the higher survival rates among patients with severe stroke.^[36,37] In dedicated observational cohorts, the incidence of SAP ranges from 11% to 28.3% with IVT^[21,38–40] and from 19.4% to 33.5% with EVT in the setting of AIS [Table 1].

Table 1.

Studies of SAP after EVT in patients with AIS caused by LVO.

Study	Incidence (%)	Study baseline	Risk factors	Conclusions on outcomes
Liu et al[164]	21.57 (22/102)	Prospective, n = 102, pM = 64.7%, mA = 64.2 years, mNIHSS = 15	Dysphagia, total time of operation	SAP is associated with poor functional outcome and higher mortality after AIS underwent EVT
Zhu et al[165]	27.7	Retrospective, n = 112, pM = 67.9%, mA = 62 years, mNIHSS = 15	Dysphagia, higher NLR, failed recanalization	Mortality was increased 3.5 times (from 11.1% to 38.7%),
Muhl et al[166]	20	Prospective, n = 1061, pM = 46%, mA = 74 years, mNIHSS = 13	Age, female, initial NIHSS, dysphagia, prior stroke, SI-ASPECTS, MV	SAP is significantly associated with mortality and an overall poor clinical outcome.
Schaller-Paule et al[167]	19.4	Prospective, n = 4281, pM = 45.2%, mA = 78 years	Male, age, moderate NIHSS, dysphagia, unconsciousness, atrial fibrillation, diabetes mellitus	This risk of ASP is further increased if an EVT is performed
Zhang et al[168]	33.5	Retrospective, n = 248, pM = 64.2%, mA = 67.2years, mNIHSS = 13	Dysphagia, intubation, higher BMI, higher SII	Patients with AIS who develop SAP are more likely to experience negative outcomes.
Zhang et al[169]	29.58	Post hoc analysis of RCT, n = 639, pM = 56.6%, mA = 70 years, mNIHSS = 17	Heart failure, pre-thrombectomy reperfusion, higher creatinine levels, NIHSS at 24 h	SAP was associated with poor 90 -day functional outcomes in patients with AIS treated with EVT.

AIS: Acute ischemic stroke; BMI: Body mass index; EVT: Endovascular treatment; LVO: Large vessel occlusion; mA: Mean age; mNIHSS: Mean National institutes of health stroke scale; MV: Mechanical ventilation; NLR: Neutrophil–to–lymphocyte ratio; pM: Percentage of male; RCT: Randomized controlled trial; SAP: Stroke-associated pneumonia; SI-ASPECTS: Source image Alberta stroke program early CT score (only for anterior circulation strokes); SII: Systemic inflammation index.

Despite the positive impact of EVT on clinical outcomes, overall SAP rates in clinical trials were relatively lower and comparable between IVT and EVT. In MR CLEAN, the incidence of SAP was 11% after EVT and 15% after IVT.^[11] Similarly, ESCAPE reported SAP rates of 4.2% for EVT and 6% for IVT,^[13] while REVASCAT found respective rates of 14.6% and 8.7%.^[14] Of note, in a study from Finland, the incidence of SAP reached 40.5% in 380 consecutive stroke patients who underwent EVT between 2015 and 2019.^[41]

However, IVT and EVT may have the opposite effect on the incident of SAP. IVT has been associated with a reduced risk of SAP, whereas EVT may contribute to an increased risk. A cross-sectional study in the US found a significantly reduced incidence of SAP in patients who underwent IVT between 2007 and 2019. Admissions with IVT had a 20% lower risk of pneumonia after adjustments (prevalence ratio 0.80, 95% confidence interval 0.73–0.88).^[42] Furthermore, a meta-analysis suggested that the decreasing prevalence of pneumonia over time may be attributed in part to increased use of these acute recanalization therapies and improved screening for dysphagia.^[43]

In clinical practice, we have observed that the risk of SAP increases significantly when EVT is performed in stroke patients with LVO. In a prospective, population-based registry study from Germany, both EVT alone (adjusted odds ratio [OR], 1.49; 95% CI, 1.10–2.01) and combined IVT plus EVT (adjusted OR, 1.48; 95% CI, 1.37–1.61) were independently associated with increased odds of SAP among patients with LVO.^[31] While only 13.5% of patients with LVO receiving conservative treatment developed SAP, this risk nearly doubled to 25.9% with the use of EVT, primarily because of the need for intensive care, mechanical ventilation, and general anesthesia (GA) with endotracheal intubation. These reasons will be discussed in detail in the following sections. Further large prospective cohort studies are needed to clarify the independent associations of IVT, EVT, and bridging therapy with SAP risk, while accounting for stroke severity, dysphagia, general anesthesia, endotracheal intubation, and mechanical ventilation.

Mechanisms

Several mechanisms contribute to the development of SAP, including aspiration, increased susceptibility to infection as a result of stroke-induced immunodepression (SIID), and migration and spread of the gut microbiota [Figure 1].

Figure 1.

Mechanism of SAP in patients with AIS. AIS causes neurological impairment leading to dysphagia and impaired consciousness, which increases the risk of aspiration. It also triggers the release of neurotransmitters and hormones from the sympathetic nervous system, HPA axis, and parasympathetic nervous system, leading to atrophy and immunosuppression of immune organs and increasing the risk of infection. Lung damage is manifested by altered breathing patterns, pulmonary edema, inflammation, and reduced phagocytic ability of alveolar macrophages. Degradation of the intestinal barrier and dysbiosis contribute to systemic inflammation and further increase the risk of SAP. Figure generated by BioRender.com. AIS: Acute ischemic stroke; HPA: Hypothalamic–pituitary–adrenal; SAP: Stroke-associated pneumonia.

Dysphagia and aspiration

Early theories regarding the etiology of SAP emphasized stroke-induced aspiration as the primary factor.^[32,44] Stroke patients often suffer neurological injuries that weaken their swallowing and coughing reflexes, which significantly increase the risk of aspiration and, in turn, SAP.^[45] A review in 2006 found that patients with dysphagia had a 3.17-fold increase in risk of developing pneumonia and that this risk increased to 11.56-fold in patients with confirmed aspiration.^[46] Ischemic lesions in supratentorial brain regions are associated with an increased risk of aspiration, while lesions in the brainstem, particularly in the lateral medulla oblongata, typically cause oropharyngeal dysphagia.^[47]

However, the pathophysiological mechanisms of post-stroke dysphagia are complex and unlikely to be determined solely by the size and location of ischemic lesions in the central nervous system. Dysphagia could theoretically be associated with abnormal dopamine transmission. In patients with AIS, particularly those with bilateral basal ganglia infarctions, disruption of nigrostriatal dopaminergic signaling may reduce substance P levels, thereby impairing swallowing and protective airway reflexes.^[48,49] Stroke patients with dysphagia experience impairments in both the neuromuscular process of swallowing and pharyngeal sensation. This impairment is accompanied by degeneration of sensory nerve endings that release substance P, resulting in reduced levels of this neuropeptide.^[50] A Japanese study found that in patients with acute stroke, lower substance P concentrations in saliva were associated with reduced swallowing frequency and a significantly higher risk of pneumonia.^[51] More recent research has confirmed that capsaicin, which mediates the release of substance P from sensory nerve endings, can be used to treat dysphagia and reduce the risk of SAP.^[52] This topic is discussed in more detail in a later section.

Recent studies have increasingly focused on the association between age-related sarcopenia and changes in swallowing function.^[53,54] A study in which swallowing was examined in patients by flexible endoscopy following mechanical thrombectomy found that a reduction in mass of the swallowing muscles was independently associated with greater severity of early dysphagia after stroke and an increased likelihood of moderate to severe pharyngeal debris.^[55]

Although the claim contributes significantly to the risk, it does not fully explain the development of SAP.^[56] The immune system plays a crucial role in brain damage caused by ischemia. The injured brain, in turn, triggers an immunosuppressive effect that increases the risk of serious infection, which can be fatal in stroke patients.

SIID

Experimental and clinical evidence confirms that patients who have experienced a stroke, especially a severe stroke, undergo immunosuppression. SIID starts immediately after an ischemic stroke with the initiation of a neuroinflammatory process in the brain. This process triggers systemic immunodepression as a result of the excessive activation of three primary systems, namely, the sympathetic nervous system, the parasympathetic nervous system, and the hypothalamic–pituitary–adrenal axis.^[57]

After a stroke, the sympathetic nervous system becomes overactivated, leading to immunodepression via the β-arrestin2–nuclear factor-κB^[58] or the cyclic adenosine monophosphate–protein kinase A–NF-κB (cAMP-PKA-NF-κB)^[59] pathway. The primary lymphoid organs (bone marrow and thymus) and secondary lymphoid organs (spleen and lymph nodes) are abundantly innervated by autonomic efferent fibers, especially those in the sympathetic nervous system. Catecholamines are released into the bloodstream, and persistently elevated catecholamine levels can cause lymphopenia, impair monocyte function and Th1 cell activity, and increase apoptosis of lymphocytes. The hypothalamic–pituitary–adrenal axis also plays an important role in SIID. Secretion of various cytokines, including interleukin (IL)-1β, tumor necrosis factor (TNF)-α, and IL-6 after stroke, can stimulate the paraventricular nucleus of the hypothalamus to release corticotropin-releasing hormone.^[60] This, in turn, leads to excessive release of glucocorticoids, which can inhibit the production of pro-inflammatory cytokines while promoting the release of anti-inflammatory cytokines.^[61] Glucocorticoids promote apoptosis in B progenitor cells and reduce lymphopoiesis.^[62,63] The activated parasympathetic nervous system triggers release of the cholinergic neurotransmitter acetylcholine, which in turn inhibits secretion of peripheral inflammatory cytokines.^[64] After a stroke, increased cholinergic signaling suppresses the innate immune response in the lung. This suppression is mediated by the parasympathetic nerves and the α7-nicotinic acetylcholine receptor expressed on alveolar epithelial cells and macrophages,^[65] resulting in reduced production of proinflammatory cytokines, such as TNF-α. The α7-nicotinic acetylcholine receptor is considered a potential therapeutic target for the treatment of inflammatory diseases. Current research suggests that small RNA molecules may be involved in regulating the cholinergic signaling cascade after stroke by targeting cholinergic transcripts.^[66] After a stroke, the function of CD14⁺ monocytes is controlled via upregulated trasfer RNA (tRNA) fragments, such as tRNA fragment-22-WE8SPOX52, which downregulates the immune regulator Z form DNA (Z-DNA) binding protein 1.^[66] This finding has paved the way for the development of novel therapeutics and biomarkers targeting both protein and RNA levels in SAP.

The primary manifestations of SIID are lymphopenia,^[67,68] impaired shifting of activity from Th1 cells to Th2 cells,^[69,70] NK cell dysfunction,^[71] and reduced expression of human leukocyte antigen-DR (HLA-DR) on monocytes,^[69,72] resulting in decreased production of pro-inflammatory cytokines and increased production of anti-inflammatory cytokines, thereby increasing susceptibility to infection. Numerous studies have confirmed that SIID is most pronounced in the first 7 days after stroke,^[72–74] which is consistent with the typical time of onset of SAP.^[75] HLA-DR, the neutrophil–lymphocyte ratio, and other biomarkers have been shown to be hallmarks of immunodepression in critically ill patients.^[76,77] Previous studies have shown a strong association between SAP and reduced expression of HLA-DR in monocytes, which serves as a marker of immunodepression.^[72,73] The multicenter prospective PREDICT study also showed that dysphagia and SIID are independent risk factors for SAP.^[78] While the SIID concept is widely accepted in science, it has not received significant attention in clinical practice. From a clinical perspective, there are no specific signs or symptoms indicating an altered immune response, making it difficult to prospectively identify patients at risk. Furthermore, there is currently neither a clear definition for SIID nor a stratification biomarker that meets clinical needs.

Lung injury and inflammation

Previous research has mainly focused on the mechanisms of aspiration and SIID in SAP. Intrinsic changes in the lung after stroke have received relatively little attention, in contrast with the extensive research on lung damage after traumatic brain injury.^[79–81] Acute brain injury triggers massive sympathetic outflow, leading to adrenergic hypersensitivity of the pulmonary vein, which in turn leads to increased vascular permeability and pulmonary edema.^[82] Recent studies have also emphasized the significance of the brain–lung axis in the pathophysiology of SAP. Stroke disrupts the central autonomic regulation of immune function in the lung, resulting in an imbalance between pro-inflammatory and anti-inflammatory responses. This dysregulation leads to increased susceptibility to infection, local inflammation, and impaired lung function. A recent study identified the immune niche in the lungs of mice with transient middle cerebral artery (MCA) occlusion and found reductions in lymphocytes and several chemokines.^[83] This finding suggests a direct influence of AIS on the immune status of the lung. Another study that focused on lung function, histology, and the phagocytic ability of alveolar macrophages in rats after stroke found that focal ischemic stroke alters respiratory patterns, induces pulmonary edema and inflammation, and reduces the phagocytic ability of alveolar macrophages.^[84]

Moreover, neurogenic pulmonary edema caused by excessive sympathetic activation further exacerbates lung injury. Disruption of the blood–brain and blood–lung barriers allows bidirectional inflammatory signaling, amplifying systemic immune responses and contributing to pulmonary complications. This brain–lung crosstalk has also been linked with changes in fluid metabolism, whereby imbalances in fluid regulation worsen alveolar damage and predispose stroke patients to secondary lung infection. Recent research has shown that ischemic stroke in mice results in significant lung inflammation, as evidenced by increased numbers of macrophages and neutrophils in bronchoalveolar lavage fluid, but not in acute lung injury or respiratory distress syndrome.^[85] Further research found that depletion of tissue-resident macrophages alleviates lung injury after trauma or stroke rather than inhibiting the recruitment of monocyte-derived macrophages.^[86] In this mechanism of injury, circulating alarmins bind to receptor for advanced glycation end product (RAGE) receptors on lung-resident macrophages, activate the epidermal growth factor receptor (EGFR), and trigger inflammation. Protective strategies, such as blocking RAGE or EGFR, can alleviate lung damage. Targeting the RAGE–EGFR signaling pathway in tissue-resident macrophages offers a potential approach for the treatment of secondary complications of sterile tissue damage.

Gut microbiota

In clinical practice, it is common for no pathogen to be detected in stroke patients with pneumonia, and the yield from blood or sputum cultures is typically very low, suggesting that other mechanisms may be involved in the development of SAP.^[87] Intestinal bacteria have been found in the lungs of patients with SAP, but did not correspond to the bacterial spectrum usually seen in standard clinical practice.^[88] Recently, the brain–gut axis has received increasing research attention. An experiment in mice showed that stroke activates the sympathetic nervous system, leading to increased intestinal permeability and disruption of host antibacterial and intestinal barriers, facilitating the translocation and spread of potentially pathogenic bacterial strains from the intestinal microbiota.^[88] In the study by Stanley et al^[88], analyses of both spontaneously infected post-stroke mice and bacterial inoculation experiments detected bacteria in the blood and all lymph nodes examined. These findings suggest that gut-derived bacteria may reach the lungs through both hematogenous and lymphatic routes.^[88] However, these findings provide only indirect evidence for the proposed dissemination routes, and the exact routes for the translocation of gut microbiota remain unclear. Future studies should use methods, such as radiolabeling or isotope tracking, to elucidate the exact distribution routes.

The gut microbiota also plays a crucial role in protecting the host from pneumonia.^[89] It strengthens respiratory defenses by activating granulocyte–macrophage colony-stimulating factor (GM-CSF) signaling,^[89] which stimulates killing of alveolar macrophages and clearance of pathogens through extracellular signal-regulated kinase signaling. The production of pulmonary GM-CSF in response to infection is stimulated by the gut microbiota via IL-17A.^[90] Stroke leads to gut dysbiosis, which is characterized by changes in the diversity, abundance, and function of the gut microbiome.^[91] Xia et al^[92] identified intestinal dysbiosis to be an independent risk factor for SAP. Patients with AIS who develop SAP have depleted beneficial bacteria, such as Roseburia and enriched opportunistic pathogens, which almost doubles the risk of SAP, highlighting the importance of the gut microbiota in the pathogenesis of SAP. Furthermore, metabolites in the gut microbiota, such as short-chain fatty acids, are also associated with SAP. A prospective case-control study in the Netherlands that included 349 stroke patients and 51 controls found significant disturbances in the gut microbiota, including reduced numbers of butyrate-producing bacteria, which were associated with increased intestinal permeability and inflammation,^[93] further supporting the link between intestinal dysbiosis and the development of SAP. Acetate, one of these short-chain fatty acids, enhances the type 1 interferon response via activation of G protein-coupled receptor 43 (GPR43), one of the G protein-coupled receptors, resulting in a reduced viral load and inflammation in mice infected with respiratory syncytial virus, demonstrating the therapeutic potential of treatment targeted to the gut–lung axis.^[94]

Prediction

Risk factors

The risk factors for SAP have been extensively studied. Older age, male sex, stroke severity, dysphagia, atrial fibrillation, diabetes, chronic obstructive pulmonary disease (COPD), impaired consciousness, ventilation, and nasogastric tube insertion have been confirmed to be risk factors for SAP.^[95–98] In patients with AIS undergoing EVT for LVO, procedural factors, including the type of anesthesia used, contribute significantly to the risk of pneumonia. Recent meta-analyses have shown that the functional outcomes after GA with mechanical ventilation are better than those after sedation and local anesthesia.^[99–102] However, GA also significantly increases the risk of SAP. Schönenberger et al^[103] found that GA was associated with higher rates of pneumonia and complications, including hypothermia and delayed extubation. In another study, Campbell et al^[104] found no significant difference in 90-day functional outcomes between GA and conscious sedation but noted an increase in procedural complications, including pneumonia. Chabanne et al^[105] similarly observed that GA was associated with significantly higher rates of hypotension (87.4% vs. 44.9%, P <0.001) as well as a trend towards a higher incidence of pneumonia. Peng et al^[106] also found that GA was associated with a higher incidence of pneumonia, with a risk ratio of 1.43 (95% confidence interval 1.07–1.90, P = 0.01). The higher incidence of postoperative pneumonia observed in the GA group is likely attributable to intubation and ventilation and is consistent with previous findings.^[107,108] Furthermore, two observational studies identified a prolonged ventilation time after EVT to be an independent predictor of a poor outcome at 3 months and also found an association of prolonged ventilation time with higher rates of pneumonia.^[108,109] These observational studies were limited by multiple confounding factors, including infarct size and location, NIH Stroke Scale (NIHSS) score, and neurological status before intubation. Therefore, it is difficult to establish a direct relationship between the timing of extubation and the functional prognosis at 3 months. The ongoing EDESTROKE trial (www.clinicaltrial.gov NCT05847309) is evaluating the impact of early extubation (<6 h) vs. delayed extubation (at 6–12 h) on outcomes in patients with AIS who undergo EVT under GA.^[110]

There has also been some active research on biomarkers that predict the development of SAP. However, there is presently no single biomarker or group of biomarkers that are accurate enough to be translated into clinical practice, despite the possibility that biomarkers indicating changes in SIID pathways could play a significant role in the early prediction of SAP. The known biomarkers for SAP are summarized in Table 2.

Table 2.

Biomarkers of SAP and SAI.

Biomarker	Association	References
Immune response markers
WBC	Higher WBC count is an independent predictor of SAP.	[170]
NK cell	High NK cell count is associated with SAI.	[171]
Monocyte	Higher monocyte count at admission is associated with SAP in patients with AIS.	[71]
NLR	Higher NLR is associated with increased risk of SAP.	[172,173]
MLR	Higher MLR is an independent risk factor for SAP.	[174]
PLR	High PLR is an independent risk factor of SAP.	[175]
SII	Higher SII is an independent risk factor for the development of SAP in patients with stroke.	[176]
SIRI	Higher SIRI is an independent risk factor for the occurrence of SAP in AIS patients.	[177]
MHR	Higher MHR at admission is associated with SAP in patients with AIS.	[71]
IL-10	Higher IL-10 is an independent risk factor of SAI.	[178]
IL-6	Higher IL-6 is an independent predictor of SAI.	[178,179]
IL-13	Lower IL-13 is independently associated with SAI.	[180]
IL-1a	Higher plasma IL-1ra was associated with an increased risk of SAI.	[181]
TNF-γ	Decreased TNF-γ is an independent predictor of SAI	[180]
microRNA-21	Higher microRNA-21 is associated with SAI	[182]
HLA-DR	Decreased monocytic HLA-DR is associated with an increased risk of SAP.	[78]
Acute-phase reactants
CRP	Higher CRP was independently associated with the development of post-stroke infections, with the optimal time window for measurement at 24–48 h.	[122,183]
PCT	Higher PCT is an independent predictor of SAP.	[184]
SAA	Higher SAA measured at admission is an independent predictor of SAI.	[185]
LBP	Higher LBP is associated with an increased risk of SAP.	[78,178]
Stress response markers
Copeptin	Higher copeptin is an independent predictor of SAP.	[184]
MR-proADM	Higher MR-proADM is associated with SAP.	[186,187]
suPAR	Higher suPAR is associated with SAP.	[186,187]
SOD	Higher SOD is an independent predictor of SAI.	[188]
Metabolic and nutritional markers
Blood glucose	Hyperglycemia is independently associated with SAI	[189]
HDL	Higher levels of HDL cholesterol are independently associated with a lower risk of developing SAI in AIS patients.	[190]
Fibrinogen	Higher fibrinogen levels are positively correlated with SAP incidence in AIS patients.	[191]
Albumin	Hypoalbuminemia is an independent predictor of pneumonia in stroke patients.	[192,193]
Prealbumin	Lower prealbumin levels are significantly associated with an increased risk of SAI in AIS patients.	[194]
Vitamin D	Low vitamin D levels on admission are associated with a higher risk of in-hospital SAP in AIS patients.	[195]
FT3	Low FT3 levels at admission are independently associated with SAII occurrence.	[196]
Others
Gut microbiota	Gut dysbiosis is an independent risk factor of SAP.	[92,197]
HRV	HRV is an independent predictor of SAI	[198]

AIS: Acute ischemic stroke; CRP: C-reactive protein; FT3: Free triiodothyronine; HDL: High-density lipoprotein; HLA-DR: Human leukocyte antigen-DR; HRV: Heart rate variability; IL: Interleukin; LBP: Lipopolysaccharide binding protein; MHR: Monocyte to high-density lipoprotein cholesterol ratio; miRNA-21: MicroRNA-21; MLR: Monocyte to lymphocyte ratio; MR-proADM: Mid-regional proadrenomedullin; NK: Natural killer; NLR: Neutrophil to lymphocyte ratio; PCT: Procalcitonin; PLR: Platelet to lymphocyte ratio; SAA: Serum amyloid A; SAI: Stroke-associated infection; SAP: Stroke-associated pneumonia; SII: Systemic inflammation index; SIRI: Systemic inflammation response index; SOD: Superoxide dismutase; suPAR: Soluble urokinase receptor; TNF-γ: Tumor necrosis factor-gamma; WBC: White blood cell.

Studies on biomarkers for SAP have focused primarily on the immune response, inflammation, and metabolic nutrition. Biomarkers, such as the neutrophil–lymphocyte ratio, HLA-DR expression on monocytes, and the platelet-to-lymphocyte ratio, highlight the role of the immune response in SAP and reflect systemic inflammation and immune dysregulation. Inflammatory markers, such as C-reactive protein, procalcitonin (PCT), and serum amyloid A (SAA), are critical for identifying systemic inflammation and the bacterial infections that often accompany SAP. Metabolic and nutritional biomarkers, including blood glucose, high-density lipoprotein (HDL), vitamin D, and prealbumin, highlight the significant influence of metabolic health and nutritional status on susceptibility to infection. Furthermore, stress response markers, such as copeptin and mid-regional proadrenomedullin (MR-proADM), indicate activation of the hypothalamic–pituitary–adrenal axis and cardiovascular stress, which further increases the risk of SAP. However, there is no single composite biomarker that can simultaneously reflect the immune response, inflammation, and nutritional status. Therefore, integration of multiple biomarkers is being investigated to provide a more comprehensive assessment of the risk and progression of SAP.

Other recent studies have also explored imaging biomarkers for SAP, offering new avenues for early detection and intervention. Yu et al^[111] found that combining clinical data with imaging biomarkers, such as brain atrophy and core infarct volume, significantly improved the prediction of SAP in patients with AIS. Remarkably, a recent report showed that 15% of stroke patients who underwent pulmonary computed tomography (CT) on admission had imaging evidence of pulmonary infection within hours of symptom onset.^[112] Furthermore, these patients were at increased risk of presenting with clinical symptoms of pneumonia upon subsequent surveillance. In a related study, Arsava et al^[113] demonstrated that a higher visual opacity score on chest CT at admission was significantly associated with the development of SAP, with 23% of patients in their cohort developing SAP. This highlights the importance of chest CT imaging for early detection of high-risk patients and guidance on timely antibiotic interventions to improve patient outcomes. As we advance in the RVT era, integrating laboratory and imaging markers into research and clinical practice will become more important in terms of improving patient outcomes and increasing the effectiveness of preventive strategies.

Prediction models

As part of extensive research on the pathogenesis of and risk factors for SAP, several assessment tools have been developed for early prediction of the risk of SAP in clinical practice. These tools have significant predictive efficacy and clinical utility. The commonly used risk assessment scales are listed in Table 3. These seven prediction scales for post-stroke pneumonia (PSP) include data from several countries, including China, Germany, the USA, the UK, and the Republic of Korea. Except for the SDL index, which incorporates stroke history, dysphagia, and lymphocyte count, all other scales include age in the assessment. Apart from the Preventive Antibacterial Therapy in acute Ischemic Stroke (PANTHERIS) score, all other scales take the NIHSS score into account. Other commonly used variables include stroke history, dysphagia, and the baseline mRS score. The PANTHERIS and acute ischemic stroke-associated pneumonia score (AIS–APS) scores also include laboratory indicators. The prestroke Independence, Sex, Age, and National Institutes of Health Stroke Scale (ISAN) score can be used for patients with both acute ischemic and hemorrhagic stroke. In contrast, the other six scales are only applicable to patients with AIS. Among them, the PNEUMONIA score is designed to predict pneumonia within 30 days of stroke onset, and the Chumbler score targets pneumonia occurring during hospitalization. The SDL score is used to predict SAP in patients with AIS who have undergone EVT. The remaining scores specifically predict SAP, defined as pneumonia occurring within the first 7 days of stroke onset.

Table 3.

Prediction scoring systems for SAP.

Scoring system	Year and country	Stroke type	Sample size	Predictive factors	Score range	Sensitivity (%)	Specificity (%)	AUC	Limitations
The PNEUMONIA Score[199]	2006 the Republic of Korea	Ischemic	286	Age ( ≥65 years), Male sex, NIHSS ( ≥11), MV, Dysphagia	0–5	No stated	No stated	No stated	1. Small Sample Size: No internal or external validation. 2. Lenient Diagnostic Criteria: Potential misclassification of infections. 3. Uniform Scoring: Equal scores for all risk factors. 4. Exclusion of Critical Factors: Missing consciousness level and stroke location.
The Chumbler Score[200]	2010 USA	Ischemic	1363	Abnormal swallowing result, History of pneumonia, Higher NIHSS score, Found down at symptom onset, Age ( ≥70 years)	Low risk: 0, Medium risk: 1–3, High risk: ≥4	No stated	No stated	Derivation: 0.78, validation: 0.76	1. Unexplained Group Allocation: No reason for the distribution of modeling and validation groups. 2. Statistical Differences in COPD and Swallowing History: Significant differences between groups. 3. Inconsistent Swallowing Assessment: No standardized method for dysphagia assessment. 4. Calibration Reporting: No report on model calibration.
A2DS2[201]	2012 Germany	Ischemic	15,335 (derivation), 45,085 (validation)	Age ( ≥75 years), Atrial fibrillation, Dysphagia, Male sex, NIHSS (0–4, 5–15, ≥16)	0–10	91% (score ≥4)	57% (score ≥4)	0.84 (derivation), 0.84 (validation)	1. Unclear Pneumonia Diagnosis: Relies on clinician experience. 2. Non-Standard Dysphagia Assessment: No standardized tests or tools. 3. Lack of External Validation: Not validated across diverse populations
AIS-APS[202]	2012 China	Ischemic	8820 (derivation), 5882 (internal validation), 3037 (external validation)	Age, Atrial fibrillation, Congestive heart failure, COPD, Smoking, Pre-stroke dependence, NIHSS, GCS, Dysphagia, Stroke subtype, Blood glucose	0–34	No stated	No stated	0.797 (derivation), 0.785 (internal validation), 0.792 (external validation)	Complex Scoring System: Broad coverage of items, reliable for pneumonia prediction but impractical due to inclusion of three scoring items, laboratory indicators, and imaging classifications, leading to lengthy clinical assessment.
SDL[203]	2013 Germany	Ischemic (MCA territory)	223 (derivation), 112 (validation)	GCS, Age, Systolic blood pressure >200 mmHg, White blood cell count >11,000/µL	0–12	77.6% (score ≥5)	84.0% (score ≥5)	0.85 (derivation), 0.88 (validation)	1. Single-Center Study: Sample size of 335, inadequate modeling and validation groups. 2. Exclusion of Key Factors: Did not include dysphagia or consciousness impairment.
ISAN[204]	2015 UK	Ischemic, Hemorrhagic	11,551 (derivation), 11,648 (validation)	Pre-stroke Independence, Sex, Age, NIHSS	0–21	No stated	No stated	0.79 (derivation), 0.78 (validation)	Imbalance in Stroke Types: Hemorrhagic stroke only 8% of samples, potentially less accurate for predicting SAP in hemorrhagic stroke (C-statistic 0.71 [0.66–0.77]).
SDL[205]	2021 China	Ischemic (after EVT)	258 (derivation), 149 (validation)	Stroke history, Dysphagia, Lymphocyte count <1.00 × 103/µL	0–3	86.7% (SDL ≥1), 55.0% (SDL ≥2), 9.2% (SDL = 3)	42.8% (SDL ≥1), 83.3% (SDL ≥2), 99.3% (SDL = 3)	0.739 (derivation), 0.783 (validation)	1. Small Sample Size: Single-center, retrospective study; possible selection bias. 2. Specific Population: Validated in endovascular-treated stroke patients; needs evaluation in non-treated patients.

AIS: Acute ischemic stroke; AUC: Area under the curve; COPD: Chronic obstructive pulmonary disease; EVT: Endovascular treatment; GCS: Glasgow Coma Scale; MCA: Middle cerebral artery; MV: Mechanical ventilation; NIHSS: National Institutes of Health Stroke Scale.

There are still some unresolved questions about models for the prediction of pneumonia. First, the various studies use different diagnostic criteria. Previous research suggests significant differences in terminology and in the diagnostic approach to pneumonia.^[114,115] Second, few studies integrate biomarkers associated with SIID into SAP risk prediction models. However, measurement of biomarkers, such as HLA-DR, requires flow cytometric analysis within 4 h of blood sampling, which may not be available in all centers. This limitation makes standardization between laboratories challenging. Furthermore, imaging markers are crucial in the era of RVT, and studies have shown that these markers are associated with SAP.^[111,116] Further large-scale, multicenter, prospective studies that primarily use standardized and validated diagnostic criteria are needed. It is also important to capture and analyze all potential influencing factors, including laboratory and imaging markers, to ensure a comprehensive and accurate risk assessment.

Management

SAP is a common and clinically important complication after AIS and requires clear diagnostic criteria and effective strategies for treatment and prevention. Nutritional status is a critical factor in both the prevention of SAP and overall functional outcome in patients with AIS.^[117] In patients with severe dysphagia, tube feeding may be required to ensure adequate protein and calorie intake. Early tube feeding is associated with improved outcomes but does not reduce the rate of aspiration pneumonia.^[118] Currently, the clinical strategies for SAP used in stroke units include the use of broad-spectrum antibiotics once the infection has been diagnosed based on clinical criteria, along with prevention by screening for dysphagia. A systematic review of 15 studies involving patients with pneumonia complicating stroke found that aerobic gram-negative bacilli (38%) and gram-positive cocci (16%) were the pathogens most commonly isolated in positive cultures. The most frequently isolated bacterial species included Enterobacteriaceae (21.8%), with 12.8% Klebsiella pneumoniae, 10.1% Staphylococcus aureus, 9% Escherichia coli, 6% Pseudomonas aeruginosa, 4.6% Acinetobacter baumannii, and 3.5% Streptococcus pneumoniae.^[119] The reported frequency of positive culture data varied significantly between studies, ranging from 15% to 88%. This pathogen profile suggests that SAP is more similar to hospital-acquired pneumonia than community-acquired pneumonia. While appropriate sputum samples and diagnostic results can be challenging to obtain in non-ventilated stroke patients, pathogens were still identifiable, highlighting the importance of using molecular methods to improve detection. The Pneumonia in Stroke Consensus group recently recommended a standardized approach to antibiotic therapy for SAP.^[120] In this approach, empirical antibiotic treatment of SAP includes initiation of antibiotics within 4 h of suspicion or diagnosis, with emphasis on community-acquired pneumonia pathogens in early SAP and expanding to hospital-acquired pneumonia pathogens and multidrug-resistant organisms in late SAP and special circumstances. The duration of treatment should be at least 7 days and based on the patient’s clinical response.

Prevention

Two major studies aimed at preventing infections to improve stroke outcomes were published in 2015. These were the preventive antibiotics in stroke study (PASS), which included 2538 patients,^[121] and the STROKE-INF study, with 1224 patients.^[122] Their results showed that prophylactic use of empirical antimicrobial therapy in SAP did not reduce the incidence of pneumonia or improve stroke outcomes. Moreover, the more recent PRECIOUS trial, which included 1493 older stroke patients, showed no benefit of prophylactic use of metoclopramide, ceftriaxone, and paracetamol in terms of improvement or reducing the risk of infection over and above that achieved by standard care.^[123] Currently, preventive antibiotic therapy cannot be recommended for stroke patients or specific subgroups, although ongoing research continues to explore strategies for preventing SAP [Figure 2].

Figure 2.

Strategy for prevention of SAP in patients with AIS. Prevention strategies focus on three main targets. The first is management of dysphagia, which includes dysphagia screening, oral hygiene, capsaicin to stimulate the swallowing reflex, and neuromodulation therapies to improve swallowing function. The second is regulation of the gut microbiota, which includes dietary adjustments, probiotics, and fecal microbiota transplantation to restore a balanced gut–lung axis. The third is immune therapy, which involves β-blockers, RIC, and MSC therapy to modulate the immune response and reduce the risk of SAP. Figure generated by BioRender.com. AIS: Acute ischemic stroke; MSC: Mesenchymal stem cell; RIC: Remote ischemic conditioning; SAP: Stroke-associated pneumonia.

Targeting dysphagia

Screening

As previously mentioned, dysphagia is a major risk factor for SAP. Screening and treatment of dysphagia are crucial for the prevention of SAP. Dysphagia screening in stroke patients has demonstrated significant beneficial effects, including a lower incidence of pneumonia and improved functional outcomes.^[124,125] The current guidelines recommend formal dysphagia screening in patients with acute stroke shortly after admission.^[126,127] A retrospective registry study of 7171 patients found that dysphagia screening in patients with AIS significantly reduced the incidence of pneumonia, severe disability, and 1-year mortality. Patients who failed screening were at higher risk of pneumonia (adjusted OR 4.71), severe disability (adjusted OR 5.19), and mortality (adjusted hazard ratio [HR] 2.42), highlighting the importance of formalized dysphagia screening and management for improving stroke outcomes.^[128] A meta-analysis of data from 24 observational studies and six randomized controlled trials (RCTs) found that dysphagia screening significantly improved outcomes by reducing the incidence of pneumonia, mortality, length of hospital stay, and dependency on care.^[43] Commonly used dysphagia screening tools include various water swallowing tests,^[129] multi-consistency screens, such as the Gugging Swallowing Screen^[130] and volume–viscosity Swallow Test,^[131] and methods for detection of decreased spontaneous swallowing frequency or delayed response to swallowing stimuli.

Oral hygiene

Maintaining oral hygiene is an important prevention strategy for SAP. Effective oral care reduces the colonization of oral pathogens and thereby reduces the risk of aspiration pneumonia.^[132] In an RCT that included 146 patients with acute stroke and moderate or severe dysphagia, early use of the Gugging Swallowing Screen and more intensive oral hygiene significantly reduced the frequency of pneumonia.^[133] The intervention group had a pneumonia rate of 7% compared with 28% in the internal control group and 27% in the external control group. In another study that included 2771 acute stroke patients, dentist-led oral care significantly reduced the SAP rate.^[134] In that study, the SAP rate decreased from 17.6% in the pre-intervention period to 6.8% in the final phase with improved oral care, including early bedside screening within 72 h of admission. Therefore, in stroke patients with dysphagia, it is important to improve oral hygiene as much as possible using the available care resources.

Capsaicin

Stimulation of transient receptor potential channels can be an effective treatment for swallowing disorders.^[135] Capsaicin, the main component of red pepper, is an agonist of transient receptor potential vanilloid 1 (TRPV1), activating TRPV1 in peripheral sensory C fibers and stimulating the swallowing reflex.^[136] Recent studies indicate that capsaicin can effectively prevent pneumonia by improving swallowing function, enhancing the cough reflex, and increasing serum substance P levels. A randomized trial in 69 stroke patients showed that capsaicin with thermal tactile stimulation improved swallowing and reduced the incidence of pneumonia (P <0.01).^[137] In another study that included 92 patients, stimulation with capsaicin and ice was found to increase the substance P level (P = 0.007) and to reduce the cough reflex and lung problems.^[138] A study in 53 patients with hemorrhagic stroke found that nebulized capsaicin increased the cough reflex (P = 0.046) and reduced post-swallow residue (P = 0.032), further reducing the risk of pneumonia.^[139] These results suggest that capsaicin has potential as a preventive intervention in SAP, although further research in larger cohorts is needed to confirm its effectiveness and safety. Menthol^[140] and piperine^[141] also activate the TRPV1 receptor or similar receptors, and their use has been associated with improved swallowing function in patients with dysphagia of various etiologies.

In addition, transcranial direct current stimulation,^[142] repetitive transcranial magnetic stimulation,^[143] and pharyngeal electrical stimulation^[144] promote neuroplasticity, which can improve swallowing function by stimulating the swallowing network in the brain to prevent SAP.

Host-directed and immunomodulatory strategies

β-blockers

Experimental studies indicate that blocking the sympathetic nervous system or hypothalamic–pituitary–adrenal axis with β-blockers or glucocorticoid receptor antagonists may reduce SIID and infections after stroke and improve functional outcomes.^[145–147] However, the clinical relevance of these findings remains unclear. A study of 625 patients (553 with AIS and 72 with intracerebral hemorrhage) found that β-blocker therapy did not reduce the risk of pneumonia after stroke but significantly reduced the risk of urinary tract infection and was associated with higher 30-day mortality.^[148] Another study in 5212 patients, of whom 1155 were treated with β-blockers before stroke onset, and 244 received β-blockers during the acute phase, found that use of these agents during stroke was associated with reduced mortality and a lower incidence of PSP.^[149] Two ongoing clinical trials are examining the effects of propranolol in stroke patients. The first is evaluating the safety and efficacy of propranolol alone and in combination with antibiotics for reducing SAP while investigating the underlying immune mechanisms (POSITION www.clinicaltrial.gov NCT05375240). The second is a multicenter, randomized, open-label trial that is evaluating the effectiveness of propranolol in reduction of SAP in 168 patients with intracerebral hemorrhage by comparing the 7-day risk of pneumonia between propranolol and standard treatment (PRO-CHASE www.clinicaltrial.gov NCT05419193).

Remote ischemic conditioning

Remote ischemic conditioning (RIC), which is achieved by short periods of restricted blood flow and reperfusion in the arm, has been shown to be effective in protecting remote organs in both preclinical and clinical studies. A comprehensive meta-analysis of RIC for AIS, which included 17 RCTs involving 6392 patients, found that RIC significantly reduced the risk of stroke recurrence and improved the prognosis without causing additional adverse events, with the exception of benign RIC-related effects.^[150] The exact mechanism of RIC remains unclear. A recent study demonstrated that RIC attenuates oxidative stress and inflammation via activation of the nuclear factor erythroid 2-related factor 2/heme oxygenase-1 (Nrf2/HO-1) pathway, potentially improving outcomes in AIS.^[151] A randomized controlled pilot study found that RIC was a safe method for the prevention of SAP in patients with AIS, with a lower incidence of SAP in the RIC group (10.5%) than in the control group (27.3%). However, the sample size was small, and the between-group difference in incidence was not statistically significant (P = 0.249), highlighting the need for further research on its efficacy.^[152] A large clinical trial is currently evaluating the efficacy and safety of RIC in the prevention of SAP (RICA-2 www.clinicaltrial.gov NCT05982015).

Mesenchymal stem cells

The immunomodulatory functions of mesenchymal stem cells (MSCs) are well documented.^[153–155] Treatment with MSCs has been shown to improve outcomes in patients with lung diseases, including pneumonia, by immunomodulation, anti-inflammatory effects, alveolar cell regeneration, and antimicrobial properties.^[156–158] In a recent study, MSCs showed significant potential in the prevention of SAP through various mechanisms. Bone marrow-derived MSCs were found to enhance phagocytosis of lung macrophages by releasing migrasomes containing the antibacterial peptide dermcidin. These migrasomes facilitate microtubule-associated protein 1 light chain 3 (LC3)-associated phagocytosis of macrophages, improve bacterial clearance in the lungs, and reduce bacterial load.^[159] Consequently, bone marrow-derived MSCs not only have an antimicrobial effect but also modulate immune responses, making them a promising therapeutic strategy for the prevention of SAP.

Strengthening peripheral immunity could be an effective strategy to combat post-stroke infections and attenuate the systemic effects of neurogenic insults. Animal studies show that SAP can be prevented through mechanisms that include modulation of cluster of differentiation 200 (CD200)–cluster of differentiation 200 receptor 1 (CD200R1) signaling to prevent bacterial infections,^[160] inhibition of activation of the sympathetic nervous system in combination with Th1 cells, stimulation by interferon-gamma to reverse immunosuppression,^[161] and targeting neurogenic pathways to maintain the competence of NK cells.^[147] These mechanisms ultimately enhance the immune response and reduce the risk of infection but require further investigation in clinical trials.

Targeting the gut microbiota

As previously mentioned, the finding that post-stroke gut dysbiosis is associated with SAP provides a critical basis for the development of targeted microbiome interventions.^[92,162] However, there is currently a lack of research in this direction. Future studies should focus on modulating the microbiome to improve stroke outcomes using approaches such as dietary changes, antibiotic therapy, probiotics, and fecal microbiota transplantation.^[163]

Perspectives

Despite significant advances in the understanding of SAP, several challenges remain that require further investigation. First, the multifactorial mechanisms of SAP, including stroke-induced immunosuppression, dysphagia, and dysfunction of the gut–lung axis, require a more integrated approach for the study of the brain–lung interaction. Investigating the interplay between immunosuppression, systemic inflammation, and lung injury at the cellular and molecular levels could reveal new therapeutic targets. Second, although predictive models, such as the Age, Atrial fibrillation, Dysphagia, Sex, and Stroke Severity (A2DS2) and ISAN scores, are promising, they are primarily based on clinical parameters. Future research should focus on incorporating advanced biomarkers and high-resolution imaging techniques to improve risk stratification and early diagnosis. Multicenter validation of these models is essential to ensure generalizability across diverse patient populations. Third, the role of new therapies, including immunomodulation and interventions in the gut microbiota, requires thorough investigation in large-scale, high-quality RCTs. Exploring precision medicine approaches tailored to individual immune profiles could revolutionize the prevention and treatment of SAP. Finally, a standardized, multidisciplinary framework for the management of SAP is urgently needed. Integrating predictive tools, prevention strategies, and individualized treatment into clinical workflows optimizes patient outcomes. Addressing these challenges through collaborative research efforts and innovative methods will be critical to minimizing the incidence of SAP and improving long-term recovery post-stroke.

Conclusion

The advent of RVT has markedly improved the treatment of AIS. However, SAP remains a significant complication that worsens morbidity and mortality. Understanding the multiple mechanisms driving SAP, including immunodepression, dysphagia, lung injury, and changes in the gut microbiota, is critical to the development of effective prevention and treatment strategies. Current prevention measures, while useful, require further refinement and validation. The integration of advanced biomarkers and imaging techniques holds the potential for early detection of SAP and intervention, and ultimately improvement of clinical outcomes. Continued research is essential to address these challenges and optimize care for stroke patients in the RVT era.

Acknowledgments

We thank Biorender (https://www.biorender.com/) for assistance with the illustration.

Conflicts of interest

None.

Funding

This work was supported by the National Key Research and Development Program of China (No. 2018YFC1312402).