Post-extubation dysphagia in the ICU−a narrative review: epidemiology, mechanisms and clinical management (Update 2025)

Abstract

Dysphagia (i.e. an impairment in swallowing function that impacts on safety or efficiency) is present in many intensive care unit (ICU) survivors, in particular following extubation (“post-extubation dysphagia”, PED). Despite the fact that pathomechanisms leading to PED are currently incompletely understood, local as well as central neurological and neuromuscular dysfunctions may be key to development of PED. Data from prospective large-scale clinical investigations with systematic screening demonstrate that PED affects about one out of five (about 20%) of mixed medical-surgical unplanned (emergency) ICU admissions. PED is associated with an increased risk for aspiration, aspiration-induced pneumonia, malnutrition, increased ICU resource use, decreased quality of life, prolonged ICU- and hospital length of stay and increased overall morbidity and mortality. Data demonstrate that PED is an independent predictor of 90-day mortality with increased risk of death up to about one year after ICU admission. PED may be a somewhat overlooked medical problem since in many ICUs, PED is currently not routinely screened for in all patients at risk (i.e. all ICU patients) following extubation. In this review, we update the available data on PED with a focus on epidemiology, risk factors, potential aetiology and treatment approaches, as well as clinical management on ICUs.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13054-025-05492-7.

Background

Dysphagia is defined as impairment in swallowing function resulting in the difficulty or inability to transfer liquids and/or food from the mouth to the gastro-intestinal system [1–5]. Despite the clinical observation that swallowing disorders are common in critically ill patients, the incidence of dysphagia in critically ill patients was not clear until very recently. Previously, an incidence of dysphagia in ICU patients of 3–62% was assumed [6] with some authors indicating prevalence rates of up to 93% [4]. Data derived from mostly smaller analyses in heterogeneous ICU populations without systematic dysphagia screening. The initially reported ranges (3–62%) might thus reflect a rather high degree of initial uncertainty. Further, methodology of dysphagia assessment, timing of dysphagia assessment, participant recruitment methods (systematic vs. non-systematic screening) or duration of mechanical ventilation partly differed [2, 6]. Available data from a large prospective analysis with systematic screening of all patients at risk (i.e. all ICU patients) now demonstrate an incidence of about 18% in emergency-admitted adult mixed medical-surgical ICU patients [3]. The incidence of post-extubation dysphagia (PED) in elective postoperative patients was shown about 5% [3].

During recent years, further prospective analyses provided new insights into risk factors and long-term effects associated with post-extubation dysphagia in critical illness. Respective data address risk factors and comorbidities associated with PED including age, gender, BMI, sepsis, diabetes, chronic kidney disease, hypertension, COPD, myocardial infarction, advanced heart failure (NYHA class > II), transesophageal echocardiography, previous stroke, tracheostomy or endotracheal tube size, presence of gastroesophageal reflux, increased disease severity (APACHE-II and SOFA scores) and smoking [4]. Nonetheless, most of the data again derive from smaller and heterogeneous ICU cohorts and varying methodological approaches or study protocols, in particular regarding screening and/or recruitment, are used.

Importantly, duration of intubation and mechanical ventilation is now considered an important risk factor for dysphagia [1, 4, 5], with one large-scale prospective analysis indicating that one additional day of mechanical ventilation increases the risk for PED by about 20% [5]. Additionally, presence of pre-existing dysphagia (and consequently reduced functional reserve emerging from critical illness) [4], local malignancy or post-surgical medical conditions affecting anatomic structures of the swallowing tract are considered relevant risk factors [7]. A meta-analysis published in 2022 [8] did not demonstrate a single risk factor statistically significant. This was mostly attributed to a heterogeneity in swallowing assessment methods, populations, and specific aspects on timing of investigations.

Complications associated with dysphagia are frequently observed clinically on intensive care units (ICU). Clinical complications include aspiration of liquid or solid food to the respiratory tract with subsequent aspiration-induced pneumonia and/or need for (prolonged) mechanical ventilation [3, 6]. Delayed feeding (leading potentially to malnutrition and cachexia) may develop, which is associated with prolonged ICU length of stay or hospital stay, increased ICU readmission rates, and increased ICU resource use [3]. The clinical relevance of post extubation dysphagia in critically ill patients seems further underlined by the fact that dysphagia is an independent risk factor for increased 28-day (excess mortality about + 9%) and 90-day mortality in both neurological and non-neurological ICU patients [3]. Emerging data indicates that dysphagia affects long-term outcomes [9], e.g. in patients with pulmonary complications including acute respiratory distress syndrome (ARDS) [9, 10].

Due to the clinical relevance of post-extubation dysphagia for ICU practitioners, considerable efforts were made to explore current clinical practice standards and to raise knowledge and awareness on dysphagia. This seems important since a large international cross‑sectional survey including 746 ICUs (26 countries) [11] revealed that 67% of ICUs did not have a protocol to assess PED although most ICUs considered such a protocol important.

In the review presented here, we thus aim to update current knowledge with a focus on epidemiology, pathophysiology and etiology, diagnostic and therapeutic approaches, and (future) clinical management.

Methods

A literature search was performed in PubMed using a Boolean logic that applied the following search terms: “dysphagia” “swallowing dysfunction” or “deglutition disorder” and terms reflecting “critical illness” in the titles and excluding malignancies (dysphagia[Title] OR swallowing dysfunction[Title] OR swallowing impair*[Title] OR swallowing disord*[Title] OR deglutition dysfunction[Title] OR deglutition disord* OR deglutition impair*) AND (ICU[Title] OR critical illness[Title] OR intensive care[Title] OR critical care[Title] OR critically ill[Title] OR intubation[Title] OR post-extubation[Title])NOT carcin*[Title] NOT malign*[Title] NOT cancer*[Title] NOT Tumor*[Title] NOT neopla*[Title] NOT palliat*[Title]).

A time filter was applied, including studies published between 2018 (Update) to January 2025. The search revealed a total of 143 publications. Of these 143, 49 publications were excluded (n = 17 not concerning/not focused on dysphagia in critical illness, n = 6 neonatal/pediatric, n = 5 focused on anesthesia, n = 4 not available in English, n = 4 single case reports, n = 2 retracted articles/retraction note, n = 10 not fully accessible, n = 1 duplicate). A total of 94 articles were analyzed. Following this first analysis, 56 additional papers were identified by in depth analysis of in-article citations or considered relevant (n = 150 articles included).

Despite the initial search structure, the review presented here is considered a narrative review (following the search structure, 56 additional publications were identified by in-depth search within articles). A PRISMA flowchart is given (Additional file 1.).

Pathophysiology of swallowing in critical illness

Physiological swallowing involves more than 50 muscles, different cranial nerves and cortical and brainstem structures. Swallowing complexity was described in detail elsewhere with six key potential pathomechanisms leading to swallowing impairment being discussed [1, 7]. In brief, the assumed primary pathomechanisms leading to post extubation dysphagia comprise of (Fig. 1):

1. Direct laryngeal trauma: e.g. arising from trauma following intubation [7, 12]

2. Neuromuscular dysfunction: acquired neurological or neuromuscular-/muscular weakness with atrophy of the tongue and pharyngeal-laryngeal structures (also referred to as “ICU-acquired swallowing disorder”[ICU-ASD]) [7]

3. Disturbed sensory systems: triggered by dysfunctional oropharyngeal and laryngeal sensory systems due to damage to various afferent sensory fibers [12, 13]

4. Global impairment of perception: e.g. in case of delirium, critical illness, iatrogenic medication, traumatic brain injury, stroke, intracranial hemorrhage and/or inflammatory conditions [1, 7]

5. Gastroesophageal reflux: e.g. gastric emptying disorder requiring continuous tube feeding [1, 7]

6. Asynchronous breathing and swallowing due to shortening of the apnea phase [1, 7]

Fig. 1.

Pathomechanisms, risk factors and treatment of post-extubation dysphagia

Regarding pathophysiology and etiology, emerging evidence of the last years indicates that swallowing difficulties affect additional patient groups. “Presbyphagia” [14] refers to the physiological changes regarding anatomy, sensation and functional reserve in the elderly, and this might lead to functional changes of the cortical swallowing network and consecutive dysfunction of the autonomous nervous system in patients with dementia. Further, recent data from ICU-acquired weakness (ICU-AW), a potential contributing or risk factor for PED, indicates an important interplay [15].

Some authors [16] describe dysphagia at different levels: e.g. at the oropharyngeal level in neurological and neuromuscular disease, geriatric syndromes and structural abnormalities, dysphagia at the esophageal level occurs with esophagitis and reflux), with functional esophageal disorders, medication-induced effects, motility disorders and rheumatologic disease. Medication-induced dysphagia was described previously already in 2007, in particular following administration of neuroleptics [17], and increased incidence of delirium might in the future contribute to increased rates of dysphagia in the ICU.

Terminology of dysphagia on the ICU

The recent definition of dysphagia is provided by the WHO in the International Classification of Diseases (ICD), 11th revision (latest version released in 2025). It is listed under code MD93, which defines it as “Difficulty in swallowing which may result from neuromuscular disorder or mechanical obstruction […] classified into oropharyngeal dysphagia due to malfunction of the pharynx and upper oesophageal sphincter; and oesophageal dysphagia due to malfunction of the oesophagus”, with the first type more pertaining to critically ill patients. Due to the specific conditions of dysphagia in critical illness, the term “ICU-acquired swallowing disorder” was initially introduced in 2013 [7]. Nonetheless, additional terms are used to describe dysphagia in the context of critical illness, such as swallowing disorder [18] or swallowing dysfunction [19] or swallowing impairment [20] and deglutition disorder or dysfunction. A modified multi-professional Danish Delphi study concludes that “dysphagia is a functional impairment that either prevents or limits the intake of food and fluids, and which makes swallowing unsafe, inefficient, uncomfortable or affects quality of life” [21].

Epidemiology of dysphagia in critical illness

To investigate the incidence of dysphagia in critical illness, a systematic screening of all patients at risk (i.e. all ICU patients) is required. The largest prospective study with systematic dysphagia screening (“DYnAMICS”, n = 1,304 total population, n = 933 primary ICU admissions analyzed [3]) indicates a dysphagia incidence of 18.3% in emergency-admitted adult mixed medical-surgical ICU patients [3], a dysphagia screening positivity rate of about 5% in elective post-surgical ICU patients [3] and a 10.3% dysphagia positivity rate at ICU discharge.

A meta-analysis including 5,798 patients from 38 rather heterogenous and mostly small-scale clinical studies with 1,957 dysphagic episodes identified a combined weighted PED incidence of 41% (39% with silent aspiration events) [2]. Another systematic review and meta-analysis included 21 surgical studies with 6,140 patients and demonstrated a post-extubation dysphagia incidence of 43% [21]. A recent meta-analysis from 2024, including 30 studies, showed a pooled PED incidence of 36% [22]. The highest incidence of up to 93% was reported by a larger study in 2013 in a subgroup of patients with neurological conditions [23]. A recent systematic review [24] showed a prevalence of dysphagia ranging from 15 to 100% and observed that dysphagia is present in up to 74% of individuals at hospital discharge and in up to 22% of patients 10–17 months after hospital discharge.

Importantly, most available data derive from smaller investigations in distinct, partly heterogeneous, patient populations without systematic screening of all ICU patients at risk (i.e. the total ICU population). Further, bias is potentially introduced by e.g. need for written informed consent, aspects regarding ICU organization (e.g. whether mixed-medical surgical patient are investigated), inclusion of patients after an aspiration event, inclusion of patients following exams for dysphagia, late patient inclusion, and patient heterogeneity. Further, epidemiologic factors may explain some discrepancies: dysphagia positivity in critically ill patients increased during the recent COVID-19 pandemic (increase from 18,3% to 31% incidence rate in the same unit) [3, 25].

Risk factors for dysphagia in critical illness

Data from large-scale prospective studies identified the following key risk factors for PED in critical illness [5] (Fig. 1): acute neurological illness, emergency ICU admission, prolonged invasive mechanical ventilation, increased days of renal replacement therapy and increased disease severity (as indicated by increased APACHE-II scores) [5]. Recent data from ICU-AW, a potential contributing or risk factor for PED, hints to an important interplay [15]. Interestingly, evidence indicates that increased Body Mass Index (BMI) reduces the risk for dysphagia in critical illness indicating presence of an obesity “paradox” regarding post-extubation dysphagia in the critically ill [5, 25]. Overall, it appears that duration of mechanical ventilation represents one of the most important risk factors for PED in the critical care setting [26, 27].

Other recent publications strengthen the view that increased disease severity (as e.g. indicated by increased APACHE II scores) is an important risk factor for PED [26, 27] in addition to increased age [28–30] and presence of dementia [14]. Further, cardiovascular disease [28, 31], thyroid dysfunction [28], diabetes mellitus [32] and use of neuromuscular blockers (in particular in studies conducted in COVID-19 patients [31, 32]) are discussed. A study focusing on mechanically ventilated patients (duration of mechanical ventilation > 7 days) highlighted a decreased incidence of persistent dysphagia after extubation in case of reduced neuromuscular blocker use [30].

Clinical consequences of dysphagia in the critically ill

A long-term follow-up of a cohort of 933 mixed medical-surgical adult ICU patients systematically screened for dysphagia showed that dysphagia was found associated with an increased hazards for death for up to 1 year after ICU admission [9] with a 180-day mortality rate of 16% for dysphagia screening-positive patients versus 5.8% of screening-negative patients and a 360-day mortality rate of 25% for patients with versus 9.1% without dysphagia [9]. Recent studies focusing on clinical consequences show increased rates of pneumonia, longer time until enteral feeding, increased length of hospital and ICU stay and increased mortality [27, 29, 30, 33]. In dysphagia patients in general (not ICU populations) a “dysphagia handicap index” [20] was proposed and dysphagia restricts quality of eating and life in affected patients [34].

Assessment and diagnostic approaches in dysphagia

Screening

Several screening tests were proposed (Fig. 1), although most works focus on specific (selected) cohorts of patients, specific pathological conditions or a specific setting (acute or chronic, for the major part conducted in non-critical care settings), making the data heterogeneous. More recently, effort was made to propose potential “most suitable” non-instrumental assessment methods for oropharyngeal dysphagia in critical care [18]. This was done since a quick, pragmatic and precise screening test seems required which can be performed at the bedside by trained nurses [18, 35, 36].

Acute stroke patients are often screened systematically for dysphagia using existing protocols, and some authors propose to adapt these protocols to ICU patients (e.g. using the “Gugging Swallowing Screen (GUSS)-ICU” test). This test is a modified version of the original GUSS (see below for further details [37]).

A key point to consider is the ideal timing for swallowing screening. Although the “best” timing is currently unclear, consensus opinion is that it should occur early after extubation (e.g. same shift after extubation) using a sensitive screening test [18]. After this first step (screening), a confirmatory exam by a trained specialist is recommended within 24 h by international expert recommendations [18].

Water-swallow test

A pragmatic and validated screening tool is the water swallowing test (WST) that can be performed by trained ICU nursing staff. There are different variations of water swallow tests, one validated test is the GUSS-ICU [37] which consists in an indirect and direct swallow test. The first part tests six items assessing vigilance, stridor and swallowing of saliva. The second part is the swallowing test itself which is performed with four different consistencies. In contrast to the WST, this might make this test more difficult and more time-consuming to perform, training may also be more intense. When compared to FEES, the GUSS-ICU showed high sensitivity, and sufficient specificity in diagnosing dysphagia with an acceptable interrater reliability.

Further variations of the water swallow test include the bedside swallowing evaluation (BSE) [38]. Different consistencies are tested, which again makes the test less pragmatic and should be performed by an expert. Other tests include the modified Volume- Viscosity Swallow test (mV-VST) in which patients are given consecutive (three) consistencies at increasing amounts. The V-VST was initially validated in a selected patient population (stroke and amyotrophic lateral sclerosis), but a modified version of the test (mV-VST) was validated as bedside screening tool for intensive care unit patients [39].

An additional dysphagia tool specifically developed for the ICU is the Bernese ICU Dysphagia Algorithm, (BIDA) [35]. BIDA includes an initial pre-test safety check, is based on the validated WST and aims to provide a pragmatic dysphagia assessment tool that is easy to train and to perform requiring few time resources [3]. Before the WST, a “safety check” is performed in BIDA, testing e.g. that vigilance is sufficient (otherwise the test is postponed) [35]. This second step, the WST trial itself, consists in a simple swallowing assessment carried out via administration of a progressive amount of liquids (e.g. a teaspoon is given three times and then a glass of water as proposed in literature) [35]. Assessment of whether the test is passed or not is based on changes in the voice and respiratory symptoms, such as coughing, choking, breathlessness, wet or gurgly voice, or water leaking out of mouth [35]. Importantly, while the above mentioned GUSS-ICU test is formally validated, the BIDA awaits formal validation against FEES (i.e. Flexible Endoscopic Evaluation of Swallowing).

Validation of non-instrumental screening and assessment tools

Initially, a systematic review and meta-analysis assessing the accuracy of dysphagia screening using WST highlighted considerable variation in methodologic rigor, especially in reporting time from WST administration and comparison test, time-to-test and methodological homogeneity [40]. The review reports relatively good sensitivity, especially when consecutive sips are being administered, but only a moderate specificity.

The GUSS-ICU was also recently validated against instrumental assessment (FEES) showing a sensitivity of more than 90%, a good specifity, a positive predictive value of over 90% and a negative predictive value superior to 70% [37].

When comparing BSE to FEES, variable accuracy to predict aspiration in critically ill patients within the first 24 h after extubation was noted (sensitivity 77%, specificity of 65%) [38]. Nevertheless, it was recently included in the development of a decisional algorithm to detect aspiration in critically ill [41]. Another validation study for BSE against instrumental evaluation (FEES) was recently published. BSE showed an accuracy of 52%, a sensitivity of 83% and a negative predictive value of 81% in detecting aspiration. The same group created a BSE-based decision algorithm with five variables that allowed to increase accuracy (81%), sensitivity (95%) and negative predictive value (97%) of the test [42].

Instrumental tools and confirmation of diagnosis

The diagnostic confirmation of dysphagia, as stated above, requires instrumental assessment. Classical screening tools may fail to detect silent aspiration, which is then revealed only in cases of clinical consequence (i.e. aspiration-induced pneumonia/pneumonitis).

Confirming the diagnosis of PED in critically ill patients is performed using two different approaches: 1) Videofluoroscopic swallowing examination (VFSS) or 2) FEES. Additional instrumental measures addressed in the literature include ultrasonography, tissue Doppler imaging, high-resolution manometry and oropharyngo-esophageal scintigraphy (OPES), currently judged not reliable or feasible in the ICU context [1].

VFSS, also termed “modified barium swallow”, investigates all four stages of swallowing and can reveal evidence for intra-deglutitive aspiration, which is not visible in FEES (“white out”-effect due to velum elevation) [1]. The key disadvantage is that it cannot be carried out at the bedside in ICUs and it requires (resource-intense) patient transport to radiological suites (with subsequent radiation exposure).

FEES can be performed at the bedside by trained examiners at any time with little equipment, low cost and is associated with low risks overall [43]. Thus, FEES is currently the diagnostic (confirmatory) gold standard in the ICU [18]. This functional examination allows grading according to the severity of disease through direct observation of the swallowing act. The result of the FEES examination is most often graded using the penetration–aspiration scale (PAS) [44]. The PAS score assesses how deep material penetrates into the airways (above, up to the level, below the vocal cords), whether material remains in the airways and how forcefully the patient can or cannot remove the material from the airway (coughing). The PAS score can be used for both the VFSS and the FEES [13, 43].

One important advantage of FEES is that it allows to assess the mechanism of dysphagia, which may allow to classify neurogenic dysphagia into different phenotypes and correlate each of these phenotypes to specific neurological diseases [45]. FEES possesses the advantage for adjustment in patient’s diet according to the result of instrumental evaluation [46]. Some authors propose an “integrated FEES report”, a comprehensive multi-step approach to contextualize findings of flexible endoscopic evaluation including specific postural interventions during FEES and immediate assessment of their effects [13]. In order to standardize the interpretation of FEES, Curtis et al. established a rating scale, referred to as “Visual Analysis of Swallowing Efficiency and Safety” (VASES) [47].

Validation of instrumental screening and assessment tools

Despite an extensive use of PAS in the field of deglutition research, a recent systematic review including studies utilizing VFSS, confirms the existence of discrepancies in its use (difference in bolus volumes and consistencies protocols, uncertainty among clinicians when ranking PAS score, PAS reported differently as either a categorical, an ordinal or an interval variable) [48]. Recent work thus aimed to address this issue in an attempt to validate and standardize the conduction of instrumental assessments for dysphagia, for example VASES [47] with the idea to improve intra- and inter-rater reliability of FEES. Although promising, this tool necessitates validation through comparison with other validated FEES rating scales.

Other work focused on standardization and validation in an attempt to adjust these scores to instrumental investigations. “Dynamic Imaging Grade of Swallowing Toxicity” (DIGEST) is a five-point ordinal scale that assesses global severity of pharyngeal dysphagia and takes into consideration two main aspects of swallowing which is safety (graded using PAS corrected by frequency and quantity of the aspiration/penetration events) and efficiency (graded on the basis of the amount of pharyngeal residue, taking into account the bolus consistency with which they occurred). DIGEST was originally developed for head and neck cancer patients [49] and was adapted and later validated using FEES (DIGEST-FEES). Further studies also promoted validation of DIGEST-FEES in other cohorts of patients, in particular patients with neurogenic dysphagia (Parkinsons Disease) [50]. However, it may be argued that pharyngeal hypoesthesia and impaired secretion management may be important mechanisms in neurogenic dysphagia as well as oral dysfunction or phenotypical aspects of dysphagia. These factors are not considered in DIGEST-FEES, which may partially limit its applicability.

Updates on therapeutic approaches to dysphagia

A meta-analysis in 2020 [12] including therapeutic interventions on dysphagia in the acute care setting (involving mainly acute stroke patients) showed that data on treatment options for dysphagia in critical care is scarce. Nine interventions, including electrical and magnetic neurostimulation and muscle strengthening treatments were identified among 22 studies with varying outcomes. Generally, three main groups of interventional approaches can be considered: Compensatory techniques (e.g. dietary texture modifications, postural changes/compensatory maneuvers), rehabilitative interventions (improvement of swallowing function) [18] and electrical stimulation at different levels (Fig. 1).

Compensatory techniques

The aim of compensatory techniques is to reduce the risk of complications in patients with dysphagia, typically aspiration-induced pneumonia. A simple technique consists in elevation of the head of the patient’s bed, improved oral hygiene with antiseptic solutions and suctioning (preventing aspiration of saliva contaminated with oral pathogens) and patient mobilization [18, 51].

Dietary interventions

Diet interventions mainly consist in the modification of food texture, such as thickening liquids and pureeing solid foods, in order to limit aspiration-induced conditions. In this regard, the “International Dysphagia Diet Standardization Initiative” (IDDSI) committee developed a dysphagia diet framework in 2017 through systematic review aiming to provide a standardized way of naming and describing modified food textures and thickened liquid consistencies [52]. However, even if dietary adjustments in the treatment of dysphagia showed acceptable results in specific patient cohorts (e.g. stroke patients [51]), evidence for an impact on hard clinical endpoints in critically ill patients is unavailable [53].

Behavioural therapy

Behavioural therapy includes postural changes of the body and head during swallowing, aiming to immediate improvement of dysphagia and optimization of bolus flow and rehabilitative swallowing exercises including muscle training targeting a long-term effect [50]. Among these treatments, respiratory mucle strength training (RMST) [54], chin-tuck against resistance, shaker exercises and swallowing exercises/maneuvers (such as the Mendelsohn maneuver) were proposed.

Respective techniques were developed for patients with delayed swallowing reflex or anatomical changes but were validated for stroke patients [50, 53, 55]. However, there is only limited evidence for critically ill patients. One study [42] described a benefit of speech therapies of various types applied at discretion by the involved physiotherapist. The study shows how such therapies favor early progression of oral intake in patients with PED, improve airway management including cough effectiveness and swallowing outcomes [42]. Furthermore, it was demonstrated that a delay in initiation of speech and language therapy may be associated with adverse outcomes as aspiration-induced pneumonia, persisting dysphagia and death [56]. A recent systematic review analyzed 10 studies (mostly randomized controlled studies) describing the efficacy of interventions on PED. These studies deal mainly with behavioral therapies which might have beneficial effects, but quality of evidence is low [57].

Rehabilitative interventions

Behavioral therapies are often compensatory in nature but special swallowing techniques and the adoption of special postures may provide therapeutic benefits [1, 51]. Thus, there is some overlap with rehabilitative interventions aiming to enhance swallowing physiology that are widely used [1, 4, 42, 57]. A good example for a therapeutic benefit is respiratory muscle strength training (RMST) which was originally more a behavioural therapy before it was observed that expiratory muscle strength training improves airway safety in swallowing [58]. Later on, it was shown beneficial in a case series with patients suffering from deconditioning associated with tissue loss [54] and might thus represent an additional treatment approach in future.

Pharmacologic interventions

Few works explore the use of pharmacological therapy. It is known from pre-clinical animal studies that topical application of transient receptor potential (TRP) agonists in peripheral swallowing-related regions helps to initiate swallowing reflexes [59, 60]. A clinical approach was based on Transient-Receptor-Potential-Vanilloid-1 (TRPV-1) sensory receptor agonists altering sensory perception in the pharynx and initiating a swallowing reflex [61].

Since capsaicin is a specific agonist to TRPV1 and mediates a dose dependent local release of substance P, it was shown that this positively influences swallowing biomechanics in healthy subjects [62]. Nonetheless, respective medical therapies are not widely used or recommended due to scarce data availability [51].

Neurostimulation

Neurostimulation therapies seem a promising innovation in the field of dysphagia rehabilitation treatments [55, 63, 64]. The aim of neurostimulation therapies is to restore the physiology of neurological structures involved in swallowing coordination and control which are damaged due to injury to central components (in case of stroke or traumatic brain injury) or peripheral components (in case of desensitization following prolonged mechanical ventilation).

In recent years, neuromuscular electrical stimulation was tested at the cranial level with different methods of non-invasive brain stimulation (NIBS) procedures aiming to directly stimulate the cortex. Among such NIBS, transcranial direct current stimulation (tDCS) applies stimulation to the pharyngeal motor cortex [65] whereas transcranial magnetic stimulation (rTMS or TMS) is considered to promote re-organization of the swallowing cortex function by generating evoked potentials through a pulsed magnetic field [66]. Both methods are currently used for dysphagia treatment in stroke patients. In addition, a combination method, paired associative stimulation (PAS), pairing a peripheral electrical with a central cortical transcranial magnetic stimulation (TMS) was investigated [67].

At the pharyngeal level, neuromuscular electric stimulation (NMES) and pharyngeal electrical stimulation (PES) [63, 64] were evaluated. Neuromuscular electrical stimulation (NMES) involves peripheral stimulation of the pharynx and swallowing muscles through transcutaneous application of currents to the cervical region in order to stimulate axonal motor or sensory nerve endings. The treatment - mostly performed as adjunct to behavioral therapies - proved benefits in stroke patients regarding decreased incidence of dysphagia and aspiration-induced pneumonia [68, 69]. PES consists in application of an electric current to the hypopharyngeal mucosa via a catheter (typically a modified nasogastric tube) providing a sensory stimulus. PES is successfully used in stroke patients [64, 70–75]. Data from the PHAST-TRAC trial with tracheostomized stroke patients showed benefit of PES regarding early decannulation [64]. As a result, PES was included in several guidelines for the treatment of neurogenic dysphagia and tracheostomy care [55]. PES was investigated in smaller analyses in ICU patients, e.g. when applied as a preventive measure before extubation [76].

PES treatment in mixed medical-surgical ICU patients is currently tested in a randomized blinded multicenter study including patients with PED [77].

Discussion

A key aspect in PED seems to increase awareness for dysphagia on the ICU with introduction of pragmatic systematic dysphagia screening programs after extubation. In most ICUs, PED is still not routinely screened for which contributes to adverse outcomes in the critically ill. Following a pragmatic systematic screening of all patients at risk for dysphagia (i.e. the total population of ICU patients post extubation, and based on recently published international expert recommendations [18], screening positive patients should then be referred to a confirmatory specialist exam via a two-step algorithm using an instrumental swallowing assessment, ideally FEES. This includes grading of dysphagia severity using outcome measurements under visualization of an instrumental assessment (such as FEES). However, which of the proposed two-step dysphagia algorithms could be used depends on local resources, patient volume and pre-existing experience.

The clinical impact of PED on patient-related outcomes was now well established. Key features of PED include increased rates of pneumonia, longer time until enteral feeding, increased resource use on the ICU, increased length of both hospital and ICU stay and increased mortality rates. Further, in the last about 5 years, it now became evident that dysphagia is a long-term medical problem with recent data demonstrating that dysphagia often persists for several months after ICU stay [24] with mortality increased to about one year after ICU admission in mixed medical-surgical ICU populations. Data show high costs for health care systems arising from prolonged hospital stay and costly ongoing (ICU and hospital) therapies [33]. Despite emerging data demonstrating the profound impact of PED on mortality, the impact on morbidity and quality of life remains under-investigated.

In the last few years, mounting data on the influence of specific risk factors for PED in the ICU became available. This seems important since early identification of patients at increased risk may allow for targeted PED screening in situations where the total ICU population cannot be screened for organizational or resource reasons. However, it should be noted that all (previously ventilated) ICU patients are at risk for PED and should thus undergo systematic screening for dysphagia. Important risk factors for PED are emergency ICU admission, increased disease severity, underlying acute neurological disease, prolonged time of invasive mechanical ventilation and need for renal replacement therapy [5]. Interestingly, recent data suggests the fact that cohorts with ICU-acquired-weakness [15] and geriatric patients [14] may be particularly vulnerable.

Although different types of treatments for dysphagia are currently applied in clinical practice (compensatory techniques, dietary interventions, behavioral therapies and rehabilitation interventions), evidence on the effectiveness of these therapies in the acute and ICU setting is increasing during the last recent years. Often, patient cooperation of ICU patients is required (especially regarding compensatory and behavioral therapies), and thus respective therapies and interventions for dysphagia may not be optimally suited for patients post critical illness. Moreover, beneficial effects of compensatory and behavioral as well as physical therapies are often not observed immediately. Thus, further testing of additional novel interventional techniques, such as pharyngeal electrical stimulation (PES), seem required. While safety and efficacy were previously demonstrated in pilot trials with mainly neurological ICU patients [76], data from larger confirmatory studies in adult mixed medical-surgical ICU populations is currently awaited [77]. In this perspective, larger interventional studies for PED in the general non-neurological ICU patient population seem warranted.

Our review has several limitations: first, our manuscript does not fulfil all criteria of a systematic review (EQUATOR guidelines), and the rather high number of additional publications retrieved by in-depth article search may indicate that we might have missed some additional publications. Second, despite international expert recommendations now being available that propose the use of specific dysphagia algorithms on the ICU (that are mostly based on the validated water-swallow-test as pragmatic screening instrument), some of the mentioned dysphagia algorithms still await formal validation. Further, it seems important to note that validating specific screening dysphagia instruments or algorithms should imply a systematic screening approach with all ICU patients at risk included in the validation analysis rather than only including a selected patient population with dysphagia-related symptoms.

Summary of novel findings on PED in the few last years

• International expert recommendations now available [18] proposing systematic dysphagia testing in all ICU patients post invasive mechanical ventilation.

• Large-scale and prospective data available on specific risk factors for PED in mixed medical-surgical (neurological and non-neurological) ICU populations.

• Better understanding that acquired neuromuscular dysfunctions (as in ICU-AW) may overlap with PED (while exact pathomechanisms leading to PED on a local pharyngeal level remain incompletely understood in critical illness). Presbyphagia and age may be important confounders for PED.

• Data on mid- and long-term PED-induced mortality effects are now available demonstrating a substantial impact on mortality up to one year after ICU admission. Data on specific underlying morbidities leading to death and data on effects of quality of life after ICU discharge is awaited.

• First clinical interventional trials ongoing that test targeted interventions (pharyngeal electrostimulation) for PED in non-neurological ICU cohorts with result awaited.

Conclusion

Dysphagia, this includes but is not limited to post extubation dysphagia (PED), is a relevant medical condition observed in many ICU patients. PED currently seems an overlooked medical problem on many ICUs since it is currently not systematically screened for in all ICU patients following extubation. In dysphagia screening-positive cases, a confirmatory specialist exam with instrumental assessment (ideally FEES) should be performed in a timely manner.

Dysphagia treatment should then be performed by an interdisciplinary team (nursing team, ICU physicians, physiotherapy and speech language therapists) with nutritional adjustment and dietary interventions, physiotherapy and speech therapy measures (including compensatory and rehabilitative management approaches), as indicated. In the future, interventional measures, such as e.g. pharyngeal electrical stimulation, might prove suitable treatments for ICU patients with post extubation dysphagia.

Author contributions

All authors co-wrote the manuscript and F.R. produced the figure. All authors reviewed the manuscript and agreed to publication.

Funding

None.

Data availability 

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable, this study is based exclusively on published literature.

Consent for publication

Not applicable, this study is based exclusively on published literature.

Competing interests

The authors declare no competing interests