Hongo et al.

Critical Care (2022) 26:98

https://doi.org/10.1186/s13054-022-03974-6

RESEARCH Open Access

Association between timing of speech

and language therapy initiation and outcomes
among post‑extubation dysphagia patients:
a multicenter retrospective cohort study
Takashi Hongo1,2, Ryohei Yamamoto3, Keibun Liu4, Takahiko Yaguchi5, Hisashi Dote6, Ryusuke Saito6,
Tomoyuki Masuyama7, Kosuke Nakatsuka8, Shinichi Watanabe9, Takahiro Kanaya10, Tomoya Yamaguchi11,
Tetsuya Yumoto2*, Hiromichi Naito2 and Atsunori Nakao2

Abstract
Background: Post-extubation dysphagia (PED) is recognized as a common complication in the intensive care unit
(ICU). Speech and language therapy (SLT) can potentially help improve PED; however, the impact of the timing of SLT
initiation on persistent PED has not been well investigated. This study aimed to examine the timing of SLT initiation
and its effect on patient outcomes after extubation in the ICU.
Methods: We conducted this multicenter, retrospective, cohort study, collecting data from eight ICUs in Japan.
Patients aged ≥ 20 years with orotracheal intubation and mechanical ventilation for longer than 48 h, and those who
received SLT due to PED, defined as patients with modified water swallowing test scores of 3 or lower, were included.
The primary outcome was dysphagia at hospital discharge, defined as functional oral intake scale score < 5 or death
after extubation. Secondary outcomes included dysphagia or death at the seventh, 14th, or 28th day after extuba-
tion, aspiration pneumonia, and in-hospital mortality. Associations between the timing of SLT initiation and outcomes
were determined using multivariable logistic regression.
Results: A total of 272 patients were included. Of them, 82 (30.1%) patients exhibited dysphagia or death at hospital
discharge, and their time spans from extubation to SLT initiation were 1.0 days. The primary outcome revealed that
every day of delay in SLT initiation post-extubation was associated with dysphagia or death at hospital discharge
(adjusted odds ratio (AOR), 1.09; 95% CI, 1.02–1.18). Similarly, secondary outcomes showed associations between this
per day delay in SLT initiation and dysphagia or death at the seventh day (AOR, 1.28; 95% CI, 1.05–1.55), 14th day (AOR,
1.34; 95% CI, 1.13–1.58), or 28th day (AOR, 1.21; 95% CI, 1.07–1.36) after extubation and occurrence of aspiration pneu-
monia (AOR, 1.09; 95% CI, 1.02–1.17), while per day delay in post-extubation SLT initiation did not affect in-hospital
mortality (AOR, 1.04; 95% CI, 0.97–1.12).

*Correspondence: tyumoto@cc.okayama-u.ac.jp
2
Department of Emergency, Critical Care, and Disaster Medicine,
Okayama University Graduate School of Medicine, Dentistry,
and Pharmaceutical Sciences, 2‑5‑1 Shikata‑cho, Okayama Kita‑ku,
Okayama 700‑8558, Japan
Full list of author information is available at the end of the article

© The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the
original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or
other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line
to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this
licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecom-
mons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Hongo et al. Critical Care (2022) 26:98 Page 2 of 10

Conclusions: Delayed initiation of SLT in PED patients was associated with persistent dysphagia or death. Early initia-
tion of SLT may prevent this complication post-extubation. A randomized controlled study is needed to validate these
results.
Keywords: Post-extubation dysphagia, Speech and language therapy, Intensive care, Dysphagia, Aspiration
pneumonia

Background distributions of the timing of swallowing therapy initia-

Approximately 13–20 million critically ill patients are tion by SLPs and evaluate its impact on the swallowing
intubated in the intensive care unit (ICU) every year function of PED patients.
[1]. Post-extubation dysphagia (PED), the inability to
swallow food or liquid following extubation, is a major Methods
complication among patients in the ICU [2, 3]. An early The study was approved by the Ethics Committee of
study estimated its incidence as up to 62% of critically Okayama Saiseikai General Hospital (Approval number:
ill patients with recent prolonged intubation [3]. Impor- 210506), and it conforms to the provisions of the Dec-
tantly, a study showed that 87% of patients with PED laration of Helsinki. Patient consent was waived for all
failed to recover from dysphagia at ICU discharge, and participants enrolled in this study because of the retro-
64% of those patients had persistent swallowing disor- spective study design. This study was conducted follow-
ders at hospital discharge [4]. Even at six months, 23% of ing the Strengthening the Reporting of Observational
PED patients still experienced dysphagia [5]. Addition- Studies in Epidemiology (STROBE) guidelines.
ally, PED can negatively impact patient outcomes, leading
to longer hospital length of stay, aspiration pneumonia, Study design and setting
re-intubation, need to place feeding tubes, and higher This multicenter, retrospective, cohort study was per-
mortality rate [4, 6–8]. These findings emphasize the formed using data from the electronic medical records
importance of multidisciplinary efforts in the early recog- of patients admitted to the ICUs of five tertiary and
nition and management of PED; however, to date, limited three community hospitals in Japan from January 2017
evidence exists to treat this pathology in the ICU. to December 2020. The five tertiary hospitals included
A recent meta-analysis showed that swallowing treat- Kameda Medical Center in Chiba, Seirei Hamamatsu
ments had a significant effect on reducing the risk of General Hospital in Shizuoka, Nagoya Medical Center in
developing pneumonia in the critical care setting [9]. Aichi, Hokkaido Medical Center in Hokkaido, and Nara
However, most data were derived from stroke patients Prefecture General Medical Center in Nara. The three
treated with electrical stimulation. Speech and language community hospitals included Okayama Saiseikai Gen-
pathologists (SLPs), experts in the treatment of commu- eral Hospital in Okayama, Misato Kenwa Hospital in
nication disorders or swallowing impairments, primarily Saitama, and Okayama Rosai Hospital in Okayama. The
evaluate and manage patients with dysphagia in the ICU detailed characteristics of the institutions are shown in
[10, 11]. Although SLPs are supposed to play a critical Additional file 1.
role in treating this complication, the optimal timing of
interventions for PED patients by SLPs is still unclear. Study population
Early oral trials may cause a transient decrease in oxygen All patients aged ≥ 20 years were eligible if they required
saturation, while delay in interventions may result in less orotracheal intubation and mechanical ventilation
effectiveness and worse functional outcomes [9]. Some for longer than 48 h and were treated for dysphagia by
patients even recover without any specific treatment. SLPs because of a lower modified water swallowing test
These clinical concerns are causing a conflict between (MWST) score following extubation (described below).
recovery of swallowing function and the risk of aspiration To focus on the patients with PED, those admitted to the
in the absence of any established evidence for this issue. ICU due to cerebrovascular disease, those who required
More importantly, potential barriers to the care of PED tracheostomy, and those with preexisting dysphagia were
patients might stem from the fact that this complication excluded. Patients with treatment restrictions, ICU read-
is still under-recognized and less of a focus among ICU missions, and incomplete primary outcome data were
staff [2, 12]. also excluded. Patients’ dysphagia was determined to be
Therefore, we hypothesized that the timing of swallow- preexisting if they had a history of aspiration pneumo-
ing therapy by SLPs would affect the outcomes of PED nia, frequent episodes of cough upon liquid consump-
patients. Accordingly, this study aimed to investigate tion or food sticking in the throat, or required dietary
Hongo et al. Critical Care (2022) 26:98 Page 3 of 10

modification before hospital admission. Nurses routinely Data collection

screened post-extubation swallowing function using a We collected the following data from patients’ medical
bedside swallowing screening. Briefly, the MWST was records: demographics (age, gender, and body mass index
used to screen for dysphagia and score swallowing on a (BMI); comorbidities based on Charlson Comorbidity
scale of 1–5 by having the patient swallow 3 mL of water. Index scores, including preexisting dementia and cer-
We adopted the MWST because it has been widely used ebrovascular disease; ICU admission source (outpatient,
in Japan with its sensitivity and specificity of 0.70 and emergency department, general ward, transfer from
0.88, respectively, and demonstrated significant associa- another hospital); ICU admission category (cardiovascu-
tion with post-extubation pneumonia in cardiovascular lar, pulmonary, gastrointestinal, trauma, or other); ICU
surgery patient [13–15]. The MWST scores were defined admission type (medical or surgery); presence of sepsis
as follows: 1. no drinking with choking and/or respiratory or septic shock based on Sepsis-3 criteria; Acute Physio-
distress; 2. drinking and respiratory distress; 3. drinking logic and Chronic Health Evaluation (APACHE) II scores;
and choking and/or hoarseness without respiratory dis- Sequential Organ Failure Assessment (SOFA) scores at
tress; 4. drinking without choking and respiratory dis- ICU admission; the presence of delirium and SOFA score
tress; and 5. drinking without choking and respiratory on the day of extubation; MWST score following extuba-
distress, plus able to perform two repetitive dry swallows tion; duration of mechanical ventilation; use of vasopres-
within 30 s [15, 16]. Patients with MWST scores of 3 or sor, intra-aortic balloon pump (IABP), extracorporeal
lower, which were associated with a greater risk of aspira- membrane oxygenation (ECMO), or renal replacement
tion pneumonia, were subjected to swallowing therapy by therapy (RRT); data on enteral nutrition (EN) and par-
SLPs [15, 16]. The timing of speech and language therapy enteral nutrition (PN) and SLT (timing, frequency, swal-
(SLT) initiation was determined based upon the intensiv- lowing and/or compensatory rehabilitation); duration
ist’s or attending physician’s discretion. of oral intake after extubation; and outcomes including
functional oral intake scale (FOIS) score (see below) at
SLT and exposures hospital discharge and on the seventh, 14th, and 28th day
In this study, SLT was defined as at least one unit (equal after extubation, aspiration pneumonia after extubation,
to 20 min) of rehabilitation conducted by SLPs dur- in-hospital mortality, and length of ICU and hospital stay.
ing hospitalization [17]. The sessions included swallow-
ing and/or compensatory rehabilitation strategies, both Outcomes
of which are common approaches to treating dysphagia The primary outcome was dysphagia or death at hospi-
[18]. Swallowing rehabilitation involved traditional swal- tal discharge. This composite outcome was chosen based
lowing exercises to restore the strength and coordina- upon the previous studies to minimize selection bias by
tion of the tongue, pharyngeal, and laryngeal muscles, encompassing all patients who underwent SLT [20, 21].
as well as respiratory muscle strength training. Swallow- Dysphagia was defined as FOIS score < 5, which repre-
ing compensatory strategies refer to bolus modification sents a severely impaired oral intake [22, 23]. The FOIS,
(i.e., texture, volume, and consistency), environmental a validated and reliable scale for measuring dysphagia,
arrangement, or maintaining proper posture of the head, was examined by the treating SLPs. Scores range from 1
neck, and body before swallowing [9]. During hospitali- (nothing by mouth) to 7 (total oral diet with no restric-
zation, SLPs were primarily responsible for modifying tions), with higher scores indicating better swallowing
the diet, changing maneuvers, or making decisions to function [22]. Secondary outcomes included dysphagia
restrict oral intake when necessary. For those patients or death at the seventh, 14th, and 28th day after extuba-
whose swallowing function was severely impaired, direct tion, aspiration pneumonia following extubation, and in-
therapy involving swallowing texture-modified foods or hospital mortality.
liquids was suspended to avoid aspiration. In all the insti-
tutions, actual steps or orders were determined solely by Statistical analyses
clinical evaluations in the ICU and flexible endoscopic Continuous variables are presented as median and
evaluation of swallowing (FEES) or videofluoroscopic interquartile ranges (IQRs), while categorical variables
swallowing study (VFSS) was used accordingly after ICU are presented as numbers and percentages. To assess
discharge. As an adjunctive tool for swallowing evalu- the impact of timing of SLT initiation on patient out-
ation, repetitive saliva swallowing test was also used to comes, multivariable logistic regression was performed,
help determine the appropriate intervention or food adjusting for covariates including institutions, age,
modification [19]. The main exposure was the days from ICU admission type, comorbidities including preexist-
extubation to SLT initiation. These data were obtained ing dementia and cerebrovascular disease, duration of
from electronic medical records.
Hongo et al. Critical Care (2022) 26:98 Page 4 of 10

mechanical ventilation, delirium on the day of extu- Results

bation, SOFA score on the day of extubation, EN, and During the four-year period, 963 adult patients
PN. These variables were essentially selected based on (age ≥ 20 years) with MWST scores ≤ 3 following extuba-
the results of previous studies and our clinical interests tion from mechanical ventilation for more than 48 h were
[24]. The results were expressed as odds ratios (ORs) identified. After excluding cases (n = 691), 272 patients
and 95% confidence intervals (CIs). were included in our study cohort (Fig. 1).
We then performed five sensitivity analyses. (1) The demographics and clinical characteristics of the
We used alternative definitions of dysphagia as FOIS participants are listed in Table 1. The median age of the
score < 6 (total oral diet with multiple consistencies, participants was 74 years old (IQR 65–81); 188 (69.1%)
without special preparation but with specific food limi- were men and 84 (30.9%) were women. The prevalence
tations) according to the previous study [25]. (2) We of preexisting dementia and cerebrovascular disease was
excluded patients who received gastrointestinal surgery 14.3% (39/272) and 12.1% (33/272), respectively. The
to avoid possible confounding by oral intake restric- numbers of ICU admissions related to cardiovascular,
tions. (3) We excluded those who died after extuba- respiratory, and gastrointestinal disease were 34 (12.6%),
tion and received SLT because the association between 88 (32.7%), and 56 (21.9%), respectively. Of the 272
dysphagia and fatality is assumed to be bidirectional: patients included in the study, more than half had sep-
dysphagia can result from fatality. (4) To address other sis or septic shock; 114 (42.1%) underwent surgery. The
potential confounders of dysphagia, we selected differ- patients were severely ill as shown by a median SOFA
ent covariates to assess the robustness of our findings; score of 8 (IQR 6–10) and a median APACHE II score
we conducted additional multivariable logistic regres- of 23 (IQR 17–29) at ICU admission. The median dura-
sion analyses to assess the association of SLT initiation tion of mechanical ventilation was 5.3 days (IQR 3.3–7.5).
timing and patient outcomes. In Model 2, gender, age, The swallowing assessment and characteristics of SLP
BMI, SOFA score at ICU admission, sepsis, and MWST intervention are shown in Table 2. The median MWST
score were added. Model 3 incorporated the duration score on the first day of swallow screening was 3 (IQR
of mechanical ventilation, vasopressor, ECMO, IABP, 2–3). The median intervals from extubation to SLT ini-
RRT, EN, and PN. (5) We conducted an analysis divid- tiation and extubation to oral intake were one day (IQR
ing the patients into three groups based on the time 0.3–2.2) and four days (IQR 2–9), respectively. Figure 2
interval from extubation to SLT initiation to reassess describes the proportion of patients presenting with dys-
the outcomes. phagia or death at discharge based on the number of days
A p-value of < 0.05 was considered statistically signifi- from extubation to SLT initiation. Although there was a
cant. Statistical analysis was performed using Stata ver- slight increase over time in the number of patients who
sion 17 (StataCorp LP, College Station, TX). died after initiating SLT, Fig. 3 demonstrates a gradual

Fig. 1 Flowchart of our study. MV: mechanical ventilation, MWST: modified water swallowing test, ICU: intensive care unit
Hongo et al. Critical Care (2022) 26:98 Page 5 of 10

Table 1 Patient characteristics of the whole cohort improvement of swallowing function: median FOIS
Variables All (n = 272)
scores on the seventh, 14th, and 28th days after extuba-
tion and at hospital discharge were 4 (IQR 1–5), 4 (IQR
Clinical information 2–6), 5 (IQR 3–6), and 6 (IQR 5–7), respectively. The
Male, gender, n (%) 188 (69.1) prevalence of dysphagia or death at hospital discharge
Age–median [IQR], years 74 [65–81] was 30.1% (82/272) (Table 3). Table 3 demonstrates the
BMI–median [IQR], kg/m2 21.0 [18.4–23.9] association of SLT initiation timing and outcomes. After
Charlson comorbidity index–median [IQR] 2 [1–3] adjusting for covariates (i.e., institutions, age, ICU admis-
Dementia, n (%) 39 (14.3) sion type, comorbidities including preexisting dementia
Cerebrovascular disease, n (%) 33 (12.1) and cerebrovascular disease, duration of mechanical ven-
ICU admission from tilation, delirium on the day of extubation, SOFA score
Outpatients, n (%) 19 (7.0) on the day of extubation, EN, and PN), the primary out-
Emergency department, n (%) 173 (63.6) come revealed that every day of delay in SLT initiation
General ward, n (%) 65 (23.9) after extubation was associated with dysphagia or death
Transfer from another hospital, n (%) 15 (5.5) at hospital discharge (adjusted odds ratio (AOR), 1.09;
ICU admission typea 95% CI, 1.02–1.18). Similar results were obtained at dif-
Medical, n (%) 157 (57.9) ferent time points: on the seventh day (AOR, 1.28; 95%
Surgery, n (%) 114 (42.1) CI, 1.05–1.55), 14th day (AOR, 1.34; 95% CI, 1.13–1.58),
Elective operation, n (%) 17 (6.3) and 28th day (AOR, 1.21; 95% CI, 1.07–1.36) after extu-
Emergency operation, n (%) 97 (35.8) bation. In addition, the timing of SLT initiation (per day
ICU admission categoryb delay) was associated with the occurrence of aspiration
Cardiovascular, n (%) 34 (12.6) pneumonia (AOR, 1.09; 95% CI, 1.02–1.17), but it was
Respiratory, n (%) 88 (32.7) not associated with in-hospital mortality (AOR, 1.04;
Gastrointestinal, n (%) 56 (21.9) 95% CI, 0.97–1.12).
Trauma, n (%) 11 (4.1) In the sensitivity analysis, our results were unaffected
Other, n (%) 79 (29.3) by alternating the definition of dysphagia (Additional
Sepsis at ICU admission file 2), excluding the patients who had gastrointestinal
Sepsis, n (%) 146 (53.7) surgery (Additional file 3), excluding those who died
Septic shock, n (%) 136 (50.0) after SLT (Additional file 4), or developing other models
APACHE II score at ICU admission–median [IQR] 23 [17–29] using different covariates (Additional file 5). Similarly,
SOFA score at ICU admission–median [IQR] 8 [6–10] those who started SLT within one day after extubation
Endotracheal intubation duration–median [IQR], days 5.3 [3.3–7.5] were associated with a decrease in dysphagia or death at
Organ support use hospital discharge and aspiration pneumonia (Additional
Vasopressor, n (%) 212 (77.9) file 6).
ECMO, n (%) 6 (2.2)
IABP, n (%) 13 (4.8)
RRT, n (%) 83 (30.5) Discussion
Enteral nutrition, n (%) 219 (80.5) In this multicenter retrospective cohort study conducted
Enteral nutrition duration–median [IQR], days 10 [5–19] on data from patients treated in the ICUs of eight hos-
Parenteral nutrition, n (%) 73 (26.8) pitals in Japan, we evaluated the timing of SLT initiation
Parenteral nutrition duration–median [IQR], days 8 [4–15] after extubation and dysphagia or death. We found that
SOFA score on extubation day–median [IQR] 4 [2–6] more than half of the patients with PED were provided
Delirium on extubation day, n (%) 139 (51.1) SLT within 24 h after extubation. A delay in initiation
Length of ICU stay–median [IQR], days 8 [5–11] of SLT post-extubation was associated with dysphagia
Length of hospital stay–median [IQR], days 44 [27–67] or death at hospital discharge. Our findings suggest that
IQR: Interquartile range, BMI: body mass index, ICU: intensive care unit, APACHE:
initiation of SLT within 24 h after extubation in patients
acute physiologic and chronic health evaluation, SOFA: Sequential Organ Failure with swallowing disorders might be preferable in terms of
Assessment, ECMO: extracorporeal membrane oxygenation, IABP: intra-aortic preventing this important complication in the ICU.
balloon pump, RRT: renal replacement therapy
a While “disability-free survival” can be used as a valid
Of 272 patients, one was missing. bOf 272 patients, three were missing
patient-centered endpoint in perioperative outcomes
research, it can be equally useful for shared decision
making, quality metrics, and benchmarking quality of
care. Disability-free survival is a combination of survival
Hongo et al. Critical Care (2022) 26:98 Page 6 of 10

Table 2 Swallow screening test and characteristics of SLT

Variables All (n = 272)

Swallow screening
MWST score–median [IQR] 3 (2–3)
Speech and language therapy
Time interval from extubation to SLT initiation–median [IQR], days 1.0 (0.3–2.2)
Time interval from extubation to oral intake–median [IQR], ­daysa 4 (2–9)
Frequency of SLT–median [IQR], days per week 4 (3–5)
Swallowing rehabilitation, n (%) 259 (95.2)
Compensatory rehabilitation, n (%) 247 (90.8)
SLT: Speech and language therapy, MWST: modified water swallowing test, IQR: interquartile range
a
Of 272 patients, 29 were unable to resume oral intake during hospitalization

Fig. 2 Distribution of the number of patients with dysphagia or death at hospital discharge based on the time interval from extubation to SLT
initiation. SLT: speech and language therapy

(1—mortality) and a patient-reported assessment of dis- outcome can reduce selection bias. Excluding patients
ability measured with a validated questionnaire. who died after initiation of SLT prevents us from exam-
We used dysphagia or death at hospital discharge as a ining the swallowing function and the effect of SLT, thus
primary outcome. Although it can complicate the inter- resulting in selection bias. The primary outcome in this
pretation, using this composite outcome offers several study was selected based upon the rationale that the
advantages. First, this outcome allows us to evaluate dis- composite of dysphagia or death could minimize selec-
ability-free survival, which is considered a valid patient- tion bias by accounting for death as a competing risk.
centered outcome as a composite outcome of mortality Nonetheless, similar results were obtained even after
and morbidity [26]. As a matter of fact, recent studies excluding those patients who died after SLT initiation.
have evaluated disability-free survival as a composite The underlying pathophysiological mechanism of
outcome, integrating all-cause mortality, regardless of the PED is multifactorial, including direct trauma, neuro-
study design [20, 21, 27, 28]. Second, it can achieve suf- myopathy as part of ICU-acquired weakness, impaired
ficient power to detect a treatment effect. A composite oropharyngeal and laryngeal sensation, altered level of
outcome has been proposed to enhance statistical effi- consciousness, gastroesophageal reflux, and discoordi-
ciency for critical care trials [29]. Third, this composite nation of breathing and swallowing [10, 24, 30]. Given
Hongo et al. Critical Care (2022) 26:98 Page 7 of 10

Fig. 3 Distribution of FOIS scores and their trajectories over time. Early discharge from the hospital did not allow us to follow up eight and 58
patients on the 14th and 28th days after extubation, respectively. FOIS: function oral intake scale

Table 3 Multivariable logistic regression analysis of association between the timing of SLT initiation and outcomes
Outcomes All (n = 272) Unadjusted OR (95% CI) p-value Adjusted OR (95% CI) p-value

Primary outcomes
Dysphagia or death on hospital discharge, n (%) 82 (30.1) 1.12 (1.04–1.20) 0.002 1.09 (1.02–1.18) 0.009
Secondary outcomes
Dysphagia or death at the 7th day after extubation, n (%) 188 (69.1) 1.36 (1.13–1.65) 0.001 1.28 (1.05–1.55) 0.011
Dysphagia or death at the 14th day after extubation, n (%)a 137 (51.8) 1.38 (1.18–1.61) < 0.001 1.34 (1.13–1.58) < 0.001
Dysphagia or death at the 28th day after extubation, n (%)b 92 (42.9) 1.21 (1.08–1.36) 0.001 1.21 (1.07–1.36) 0.001
Aspiration pneumonia after extubation, n (%) 79 (29.0) 1.11 (1.04–1.19) 0.002 1.09 (1.02–1.17) 0.012
In-hospital mortality, n (%) 31 (11.4) 1.04 (0.98–1.11) 0.135 1.04 (0.97–1.12) 0.203
Variables for the outcomes in the multivariable logistic regression included timing of SLT initiation, institutions, age, ICU admission type, preexisting dementia,
cerebrovascular disease, duration of mechanical ventilation, delirium on the day of extubation, SOFA score on the day of extubation, EN, and PN. SLT: speech and
language therapy, CI: confidence interval, OR: odds ratio, ICU: intensive care unit, SOFA: Sequential Organ Failure Assessment, EN: enteral nutrition, PN: parenteral
nutrition
a
Of 272 patients, eight were missing
b
Of 272 patients, 58 were missing

these proposed mechanisms, early intervention against ongoing dysphagia. Moreover, a previous study reported
PED is a potential efficient strategy, similar to how early that early oral care and SLT (within 24 h after admission)
mobilization appears to prevent the development of ICU- improved swallowing function and survival after hem-
acquired weakness [31]. Although several studies have orrhagic stroke [34]. Another investigation also showed
shown a favorable efficacy of swallowing and oral care similar results, where an early intervention group (three
or speech therapy delivered by specially trained nurses days after stroke) had better swallowing performance
or SLPs on PED [25, 32], optimal timing for the initia- compared a late post-stroke intervention group [35].
tion of SLT has not been explored. Several studies have Early initiation of SLT, as shown in our study, may be a
looked at the timing of dysphagia assessment or SLT beneficial strategy to improve swallowing dysfunction
among non-PED patients. In a large prospective cohort and prevent subsequent adverse events post-extubation
study, delays in dysphagia screening were associated with in critically ill patients.
a greater risk of pneumonia in acute stroke patients [33]. Although detailed time metrics were not available,
It was speculated that this phenomenon resulted from SLT was provided later than 24 h after extubation dur-
inappropriate initiation of oral intake or delayed care for ing the ICU stay in past studies [8, 25]. In these studies,
Hongo et al. Critical Care (2022) 26:98 Page 8 of 10

more than half of the patients with dysphagia received previous prospective study showed that the FEES was
their first SLT session within 24 h following extuba- safe, well tolerated, and significantly correlated with FOIS
tion. It might be challenging to initiate SLT within this score [42]. In terms of feasibility, FEES, which can be per-
timeframe in terms of clinical feasibility and given the formed at the bedside, would be preferable to VFSS in the
expectation of recovery from dysphagia within 24–48 h ICU setting [43]. Nonetheless, the MWST was deemed
post-extubation [36–38]. Nonetheless, this strategy of a pragmatic bedside assessment without requiring addi-
early SLT initiation might have contributed to early reso- tional equipment or resources [14, 15]. Forth, daily SLT
lution of dysphagia; our results showed that swallowing was completely dependent on the local policies within
function had not been restored in 30% of PED patients the above-mentioned framework, and practice compli-
at hospital discharge, which was less than a previous ance was not examined. Therefore, standardized and
prospective observational study (58/90, 64%), despite a validated approach to care PED patients should be estab-
higher patient age and severity in our study subjects [4]. lished to ensure our findings. Finally, long-term follow-
To our knowledge, this study was the first to link the up may be needed, considering the profound impact of
timing of SLT in PED patients in the ICU with their out- PED on quality of life.
comes. According to the results from national surveys,
many ICUs have become aware of this important issue;
however, standardized therapeutic strategies to treat PED Conclusions
have not yet been established [2, 39, 40]. Furthermore, In this retrospective analysis, we found that delayed tim-
a recent large international survey revealed that SLPs ing of SLT initiation was associated with poor outcomes
were not yet even available in approximately one third of such as persistent dysphagia or death and aspiration
ICUs [41], which implies that outcome effects could be pneumonia after extubation. Our observations indicate
different from the findings of this study in the setting of that the initiation of SLT within 24 h after extubation
absence of SLPs in the ICU. Although further research in patients with PED appears to be a favorable strategy
is warranted to determine and validate proper timing of for early recovery from swallowing disorders in the ICU.
SLT and which patients to prioritize, implementing early Since this was a retrospective study, a randomized con-
initiation of SLT could be a part of a global strategy to trolled trial seems warranted to confirm these results.
reduce the incidence of persistent PED.
Abbreviations
Limitations AOR: Adjusted odds ratio; APACHE: Acute physiologic and chronic health
evaluation; BMI: Body mass index; CI: Confidence interval; ECMO: Extracorpor-
First, the timing of SLT initiation was completely depend- eal membrane oxygenation; EN: Enteral nutrition; FEES: Flexible endoscopic
ent on the physician’s preference. Why SLT was delayed evaluation of swallowing; FOIS: Functional oral intake scale; IABP: Intra-aortic
for some patients was uncertain: It might have been balloon pump; ICU: Intensive care unit; IQR: Interquartile range; MWST: Modi-
fied water swallowing test; OR: Odds ratio; PED: Post-extubation dysphagia;
attributed to patient factors and/or physicians expecting PN: Parenteral nutrition; RRT​: Renal replacement therapy; SLT: Speech and
spontaneous recovery. To minimize the bias, in addition language therapy; SLP: Speech and language pathologist; SOFA: Sequential
to institutions, SOFA scores and the presence or absence Organ Failure Assessment; VFSS: Videofluoroscopic swallowing study.
of delirium on the day of extubation, which might lead to
a delay in starting SLT, were taken into consideration in Supplementary Information
the multivariable analysis. Due to retrospective design The online version contains supplementary material available at https://​doi.​
org/​10.​1186/​s13054-​022-​03974-6.
of the study, we still must acknowledge that uncaptured
clinical data such as trajectories of actual respiratory or
mental status post-extubation may play a causative role Additional file 1: Characteristics of each institution and ICU

in SLT initiation timing. Second, this study did not col- Additional file 2: Multivariable logistic regression analysis of association
between the timing of SLT initiation and outcomes, alternating the defini-
lect data on the size of the endotracheal tube. A recent tion of dysphagia as FOIS < 6
prospective study found that larger endotracheal tube Additional file 3: Multivariable logistic regression analysis of association
size induced post-extubation aspiration in patients with between the timing of SLT initiation and outcomes, excluding those who
acute respiratory failure [35]. Related to the first limita- had gastrointestinal surgery
tion, additional unmeasured or potential cofounders Additional file 4: Multivariable logistic regression analysis of association
may exist. To overcome these limitations, a randomized between the timing of SLT initiation and outcomes, excluding fatal cases

controlled trial should be conducted. Third, all partici- Additional file 5: Multivariable logistic regression analysis of association
between the timing of SLT initiation and outcomes, adjusting for different
pants were screened for PED using the MWST; however, covariates
instrumental assessment via FEES or VFSS has become Additional file 6: Multivariable logistic regression analysis of association
implemented in recent years because these provide com- between the timing of SLT initiation and outcomes, stratifying the popula-
prehensive picture of the swallowing process [42]. A tion into three groups.
Hongo et al. Critical Care (2022) 26:98 Page 9 of 10

Acknowledgements 3. Skoretz SA, Flowers HL, Martino R. The incidence of dysphagia following
We thank the members of the Japanese Society of Education for Physicians endotracheal intubation: a systematic review. Chest. 2010;137:665–73.
and Trainees in Intensive Care-Clinical Trial Group (JSEPTIC-CTG) for suggest- 4. Schefold JC, Berger D, Zürcher P, Lensch M, Perren A, Jakob SM, et al. Dys-
ing the study concept and design. The authors thank Christine Burr for editing phagia in mechanically ventilated ICU patients (DYnAMICS): a prospective
the manuscript. observational trial. Crit Care Med. 2017;45:2061–9.
5. Brodsky MB, Huang M, Shanholtz C, Mendez-Tellez PA, Palmer JB, Colan-
Author contributions tuoni E, et al. Recovery from dysphagia symptoms after oral endotracheal
TH, RY, KL, TYu, and HN designed the concept of the study. TH, RY, TYag, HD, RS, intubation in acute respiratory distress syndrome survivors: a 5-year
TM, KN, SW, TK, TYam, and HN conducted the study and performed the data longitudinal study. Ann Am Thorac Soc. 2017;14:376–83.
acquisition. TH, RY, KL, TYu, HN, and AN assessed the quality of the study and 6. Macht M, King CJ, Wimbish T, Clark BJ, Benson AB, Burnham EL, et al. Post-
performed the analysis and interpretation. TH, RY, TYu, HN, and AN wrote the extubation dysphagia is associated with longer hospitalization in survi-
manuscript, and the other authors made substantial revisions and edits. All vors of critical illness with neurologic impairment. Crit Care. 2013;17:R119.
authors read and approved the final manuscript. 7. Macht M, Wimbish T, Clark BJ, Benson AB, Burnham EL, Williams A, et al.
Postextubation dysphagia is persistent and associated with poor out-
Funding comes in survivors of critical illness. Crit Care. 2011;15:R231.
No funding was received for this work. 8. El Gharib AZG, Berretin-Felix G, Rossoni DF, Seiji YS. Effectiveness of
therapy on post-extubation dysphagia: clinical and electromyographic
Availability of data and materials findings. Clin Med Insights Ear Nose Throat. 2019;12:117955061987336.
The datasets from this study are available from the corresponding author 9. Duncan S, McAuley DF, Walshe M, McGaughey J, Anand R, Fallis R,
upon request. et al. Interventions for oropharyngeal dysphagia in acute and criti-
cal care: a systematic review and meta-analysis. Intensive Care Med.
2020;46:1326–38.
Declarations 10. Macht M, Wimbish T, Bodine C, Moss M. ICU-acquired swallowing disor-
ders. Crit Care Med. 2013;41:2396–405.
Ethics approval and consent to participate 11. Brodsky MB, Mayfield EB, Gross RD. Clinical decision making in the ICU:
This study conforms to the principles outlined in the Declaration of Helsinki dysphagia screening, assessment, and treatment. Semin Speech Lang.
and was approved by ethics committee of the Okayama Saiseikai General 2019;40:170–87.
Hospital, ID: 210506. Patient consent was waived for all participants enrolled in 12. Fernando SM, Seely AJE. Post-extubation dysphagia: the truth is hard to
this study because of the retrospective study design. swallow. Chest. 2020;158:1806–7.
13. Horiguchi S, Suzuki Y. Screening tests in evaluating swallowing function.
Consent for publication JMAJ. 2011;54:31–4.
Not applicable. 14. Estupiñán-Artiles C, Regan J, Donnellan C. Dysphagia screening in resi-
dential care settings: A scoping review. Int J Nurs Stud. 2021;114:103813.
Competing interests 15. Oguchi N, Yamamoto S, Terashima S, Arai R, Sato M, Ikegami S, et al. The
The authors declare that they have no competing interests. modified water swallowing test score is the best predictor of postop-
erative pneumonia following extubation in cardiovascular surgery A
Author details retrospective cohort study. Medicine. 2021;100:e24478.
1
Department of Emergency, Okayama Saiseikai General Hospital, 2‑25 Kokutai- 16. Tohara H, Saitoh E, Mays KA, Kuhlemeier K, Palmer JB. Three tests
tyo, Okayama Kita‑ku, Okayama 700‑8511, Japan. 2 Department of Emergency, for predicting aspiration without videofluorography. Dysphagia.
Critical Care, and Disaster Medicine, Okayama University Graduate School 2003;18:126–34.
of Medicine, Dentistry, and Pharmaceutical Sciences, 2‑5‑1 Shikata‑cho, 17. Sakai K, Momosaki R. Real-world effectiveness of speech therapy time on
Okayama Kita‑ku, Okayama 700‑8558, Japan. 3 Department of Healthcare cognitive recovery in older patients with acute stroke. Prog Rehabil Med.
Epidemiology, School of Public Health, Graduate School of Medicine, Kyoto 2016;1:20160004.
University, Yoshida‑honmachi, Kyoto Sakyo‑ku, Kyoto 606‑8501, Japan. 4 Criti- 18. Wirth R, Dziewas R, Beck AM, Clavé P, Hamdy S, Heppner HJ, et al.
cal Care Research Group, Faculty of Medicine, University of Queensland, The Oropharyngeal dysphagia in older persons—from pathophysiology to
Prince Charles Hospital, 627 Rode Rd, Chermside, Brisbane, QLD 4032, Aus- adequate intervention: a review and summary of an international expert
tralia. 5 Department of Intensive Care Medicine, Kameda Medical Center, 929 meeting. Clin Interv Aging. 2016;11:189–208.
Higashicho, Kamogawa, Chiba 296‑0041, Japan. 6 Department of Emergency 19. Persson E, Wårdh I, Östberg P. Repetitive saliva swallowing test: norms,
and Critical Care Medicine, Seirei Hamamatsu General Hospital, 2‑12‑12 Sum- clinical relevance and the impact of saliva secretion. Dysphagia.
iyoshi, Hamamatsu Naka‑ku, Shizuoka 430‑8558, Japan. 7 Department of Emer- 2019;34:271–8.
gency, Misato Kenwa Hospital, 4‑494‑1 Takano, Misato, Saitama 341‑0035, 20. McNeil JJ, Woods RL, Nelson MR, Reid CM, Kirpach B, Wolfe R, et al. Effect
Japan. 8 Department of Anesthesiology, Okayama Rosai Hospital, 1‑10‑25 of aspirin on disability-free survival in the healthy elderly. N Engl J Med.
Chikkomidorimachi, Okayama Minami‑ku, Okayama 702‑8055, Japan. 9 Depart- 2018;379:1499–508.
ment of Rehabilitation, Nagoya Medical Center, NHO, 4‑1‑1 Sannomaru, , 21. Hodgson CL, Higgins AM, Bailey MJ, Mather AM, Beach L, Bellomo R,
Nagoya Naka‑ku, Aichi 461‑0001, Japan. 10 Department of Rehabilitation, Hok- et al. The impact of COVID-19 critical illness on new disability, functional
kaido Medical Center, NHO, 7‑1‑1 Yamanote5jo, Sapporo Nishi‑ku, Hokkaido outcomes and return to work at 6 months: a prospective cohort study.
063‑0005, Japan. 11 Division of Critical Care Medicine, Nara Prefecture General Crit Care. 2021;25:382.
Medical Center, 2‑897‑5 Shichijonishimachi, Nara, Nara 630‑8581, Japan. 22. Crary MA, Mann GD, Groher ME. Initial psychometric assessment of a
functional oral intake scale for dysphagia in stroke patients. Arch Phys
Received: 28 December 2021 Accepted: 31 March 2022 Med Rehabil. 2005;86:1516–20.
23. Galovic M, Stauber AJ, Leisi N, Krammer W, Brugger F, Vehoff J, et al.
Development and validation of a prognostic model of swallowing
recovery and enteral tube feeding after ischemic stroke. JAMA Neurol.
2019;76:561–70.
References 24. Zuercher P, Moret CS, Dziewas R, Schefold JC. Dysphagia in the intensive
1. Adhikari NK, Fowler RA, Bhagwanjee S, Rubenfeld GD. Critical care and care unit: epidemiology, mechanisms, and clinical management. Crit
the global burden of critical illness in adults. Lancet. 2010;376:1339–46. Care. 2019;23:1–11.
2. Macht M, Wimbish T, Clark BJ, Benson AB, Burnham EL, Williams A, et al. 25. Wu CP, Xu YJ, Wang TG, Ku SC, Chan DC, Lee JJ, et al. Effects of a swal-
Diagnosis and treatment of post-extubation dysphagia: results from a lowing and oral care intervention for patients following endotracheal
national survey. J Crit Care. 2012;27:578–86. extubation: a pre- and post-intervention study. Crit Care. 2019;23:350.
Hongo et al. Critical Care (2022) 26:98 Page 10 of 10

26. Lönnqvist PA. Medical research and the ethics of medical treatments:
disability-free survival. Br J Anaesth. 2017;118:286–8.
27. Zhou Z, Ofori-Asenso R, Curtis AJ, Breslin M, Wolfe R, McNeil JJ, et al.
Association of statin use with disability-free survival and cardiovascular
disease among healthy older adults. J Am Coll Cardiol. 2020;76:17–27.
28. Sun LY, Eddeen AB, Mesana TG. Disability-free survival after major cardiac
surgery: a population-based retrospective cohort study. CMAJ Open.
2021;9:E384–93.
29. Heyland DK, Muscedere J, Drover J, Jiang X, Day AG. Canadian Critical
Care Trials Group: persistent organ dysfunction plus death: a novel, com-
posite outcome measure for critical care trials. Crit Care. 2011;15:R98.
30. Brodsky MB, Pandian V, Needham DM. Post-extubation dysphagia: a prob-
lem needing multidisciplinary efforts. Intensive Care Med. 2020;46:93–6.
31. Zhang L, Hu W, Cai Z, Liu J, Wu J, Deng Y, et al. Early mobilization of
critically ill patients in the intensive care unit: a systematic review and
meta-analysis. PLoS ONE. 2019;14:1–16.
32. Turra GS, Schwartz IVD, Almeida ST, Martinez CC, Bridi M, Barreto SSM.
Efficacy of speech therapy in post-intubation patients with oropharyn-
geal dysphagia: a randomized controlled trial. Codas. 2021;33:e20190246.
33. Bray BD, Smith CJ, Cloud GC, Enderby P, James M, Paley L, et al. The
association between delays in screening for and assessing dysphagia
after acute stroke, and the risk of stroke-associated pneumonia. J Neurol
Neurosurg Psychiatry. 2017;88:25–30.
34. Takahata H, Tsutsumi K, Baba H, Nagata I, Yonekura M. Early intervention
to promote oral feeding in patients with intracerebral hemorrhage: a
retrospective cohort study. BMC Neurol. 2011;11:6.
35. Bakhtiyari J, Sarraf P, Nakhostin-Ansari N, Tafakhori A, Logemann J, Faghi-
hzadeh S, et al. Effects of early intervention of swallowing therapy on
recovery from dysphagia following stroke. Iran J Neurol. 2015;14:119–24.
36. Krisciunas GP, Langmore SE, Gomez-Taborda S, Fink D, Levitt JE, McK-
eehan J, et al. The association between endotracheal tube size and
aspiration (during flexible endoscopic evaluation of swallowing) in acute
respiratory failure survivors. Crit Care Med. 2020;48:1604–11.
37. Beduneau G, Souday V, Richard JC, Hamel JF, Carpentier D, Chretien JM,
et al. Persistent swallowing disorders after extubation in mechanically
ventilated patients in ICU: a two-center prospective study. Ann Intensive
Care. 2020;10:138.
38. Marvin S, Thibeault S, Ehlenbach WJ. Post-extubation dysphagia: does
timing of evaluation matter? Dysphagia. 2019;34:210–9.
39. Zuercher P, Moret C, Schefold JC. Dysphagia in the intensive care unit in
Switzerland (DICE) - results of a national survey on the current standard
of care. Swiss Med Wkly. 2019;149:w20111.
40. Cardinal LA, Freeman-Sanderson A, Togher L. The speech pathology
workforce in intensive care units: results from a national survey. Aust Crit
Care. 2020;33:250–8.
41. Spronk PE, Spronk LEJ, Egerod I, McGaughey J, McRae J, Rose L, et al. Dys-
phagia in intensive care evaluation (DICE): an international cross-sectional
survey. Dysphagia. 2022. https://​doi.​org/​10.​1007/​s00455-​021-​10389-y.
42. Dziewas R, Auf dem Brinke M, Birkmann U, Bräuer G, Busch K, Cerra F, et al.
Safety and clinical impact of FEES—results of the FEES-registry. Neurol
Res Pract. 2019;1:16.
43. Dziewas R, Allescher HD, Aroyo I, Bartolome G, Beilenhoff U, Bohlender J,
et al. Diagnosis and treatment of neurogenic dysphagia—S1 guideline of
the German Society of Neurology. Neurol Res Pract. 2021;3:23.

Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in pub- Ready to submit your research ? Choose BMC and benefit from:
lished maps and institutional affiliations.
• fast, convenient online submission
• thorough peer review by experienced researchers in your field
• rapid publication on acceptance
• support for research data, including large and complex data types
• gold Open Access which fosters wider collaboration and increased citations
• maximum visibility for your research: over 100M website views per year

At BMC, research is always in progress.

Learn more biomedcentral.com/submissions