INT J LANG COMMUN DISORD, SEPTEMBER–OCTOBER 2016,

VOL. 51, NO. 5, 556–567

Research Report
Patterns of return to oral intake and decannulation post-tracheostomy across
clinical populations in an acute inpatient setting
Lee Pryor†‡, Elizabeth Ward‡§, Petrea Cornwell¶, Stephanie O’Connor†# and Marianne Chapman†#
†Royal Adelaide Hospital, Adelaide, SA, Australia
‡The University of Queensland, School of Health & Rehabilitation Sciences, St Lucia, QLD, Australia
§Centre for Functioning & Health Research (CFAHR), Queensland Health, Buranda, QLD, Australia
¶The Prince Charles Hospital, Metro North Hospital and Health Service, Chermside, QLD, Australia
School of Applied Psychology, Menzies Health Institute Queensland, Griffith University, Mount Gravatt, QLD, Australia
#The University of Adelaide, School of Medicine, Adelaide, SA, Australia
(Received June 2015; accepted November 2015)

Abstract
Background: Dysphagia is often a comorbidity in patients who require a tracheostomy, yet little is known about
patterns of oral intake commencement in tracheostomized patients, or how patterns may vary depending on the
clinical population and/or reason for tracheostomy insertion.
Aims: To document patterns of clinical management around the commencement of oral intake throughout hospital
admission and along the decannulation pathway in patients with a new tracheostomy, and to examine the nature
of variability across multiple clinical populations.
Methods & Procedures: A 12-month retrospective review of 126 patients who had undergone an acute tracheostomy
was conducted. Within the cohort, patients were further classified into eight clinical populations representing
specialty areas within the tertiary referral centre. Data were collected on timing of milestones and patterns of
clinical management related to oral and enteral feeding and decannulation. Relationships between temporal
variables were calculated, in addition to descriptive analysis of the overall cohort and by clinical population.
Outcomes & Results: Median temporal markers of patient progression post-tracheostomy insertion for the cohort
were: continuous cuff deflation after 7.5 days, commencement of oral intake after 10.5 days, decannulation
after 15 days and cessation of enteral nutrition (EN) after 17 days. However, considerable individual variation
and differences between clinical populations was observed. Overall, 86% of the cohort returned to oral intake,
although 25% were discharged with EN via a gastrostomy. A total of 86% of the group were decannulated by
hospital discharge. Oral intake was introduced at every stage of the decannulation pathway, including prior to
cuff deflation, but the majority of patients commenced diet/fluids following cuff deflation or with an uncuffed
tube in situ, and most patients who ceased EN did so following decannulation. Commencement of oral intake
was evenly split between the intensive care unit (ICU) and the wards. Increased time to commencement of oral
intake correlated with increased time to decannulation (r = .805, p = .001), and increased time to decannulation
correlated with increased hospital length of stay (r = .687, p = .006). Whilst cohort patterns were observed within
the heterogeneous group, sub-analysis revealed distinct patterns of oral intake management across the different
clinical populations.
Conclusions & Implications: The data provide benchmarks enabling comparison by overall cohort as well as by
specialist clinical populations, each with differing reasons for tracheostomy insertion. The data would suggest that
tracheostomy patients should not be looked upon as a singular cohort; rather, evaluation of factors with specific
attention made to underlying aetiology and individual clinical presentation is essential.

Keywords: swallowing, dysphagia, oral intake, tracheostomy, transitional feeding, decannulation.

Address correspondence to: Lee N. Pryor, Speech Pathology Department, Royal Adelaide Hospital, North Terrace, Adelaide, SA 5000, Australia;
e-mail: lee.pryor@sa.gov.au
International Journal of Language & Communication Disorders
ISSN 1368-2822 print/ISSN 1460-6984 online  C 2016 Royal College of Speech and Language Therapists

DOI: 10.1111/1460-6984.12231
Oral intake and decannulation post-tracheostomy 557

What this paper adds?

What is already known on this subject?
Dysphagia often co-exists for patients with a tracheostomy and relates to the underlying medical reason for tra-
cheostomy insertion. Currently, there is insufficient information on the patterns of management, including return
to oral intake, for tracheostomized patients who have a variety of clinical diagnoses. Prognostic decision making for
speech–language therapists is therefore limited.
What this study adds?
The data provide initial benchmarks and highlights important differences between clinical populations and patterns
of management related to oral intake and decannulation throughout acute hospital admission.

Introduction (Amathieu et al. 2012). Furthermore, while some re-

search has suggested aspiration risk to be greater when
Tracheostomy insertion is a commonly performed pro-
swallowing with a tracheostomy cuff inflated compared
cedure in critically ill patients, with incidence increasing
with deflated (Davis et al. 2002, Ding and Logemann
due to advances in insertion techniques (Freeman et al.
2005), these studies had methodological limitations.
2000) and utilization of life-prolonging interventions in
Hence, there is no robust evidence to support a dif-
an ageing and increasingly medically complex critically
ference in swallowing according to tracheostomy pres-
ill population (Adhikari and Rubenfeld 2011). Over the
ence or absence, or as conditions change along the de-
past two decades, involvement of the speech–language
cannulation pathway. Individual variation, however, is
therapist (SLT) with tracheostomized patients has grown
commonly reported.
alongside the increases in patient numbers and critical
The impact of age may be an additional compli-
care advances. In part this has been enabled by innova-
cating factor in tracheostomy management. An older
tions in communication facilitation, and increased ac-
person with comorbidities may be vulnerable to func-
cess to instrumental assessment including fibre optic
tional decline with exposure to stressors such as critical
endoscopic evaluation of swallowing (FEES). There is
illness (Brummel et al. 2015); furthermore, we know
also greater awareness of the role of the SLT in providing
increasing age can affect the normal swallow (Daggett
ongoing assessment and recommendations to the mul-
et al. 2006; Robbins et al. 1992). Age, functional reserve
tidisciplinary team regarding upper airway function and
and number of days post-tracheostomy have therefore
airway protection, which contribute to team decisions
been suggested as important factors when assessing pa-
regarding patient suitability for cuff deflation and tube
tients for oral intake, with older tracheostomized pa-
removal (Hales et al. 2008, Warnecke et al. 2013).
tients reported to take longer to swallow successfully
Dysphagia, leading to impaired airway protection,
(Leder 2002).
is a common coexisting issue for many patients with
Greater acceptance for dysphagia to be a factor of
a tracheostomy. Early literature suggested that the pres-
the patients’ medical and physical condition, rather than
ence of a tracheostomy tube caused or increased dyspha-
tube presence, has enabled a change in clinical practice.
gia (Bonanno 1971). Current theory, however, supports
Whilst clinicians may have previously deferred swallow-
that it is the underlying aetiology or condition neces-
ing evaluation and/or oral trials until the tracheostomy
sitating tracheostomy insertion that is the major causal
could be removed, there is now greater focus on com-
factor associated with dysphagia in tracheostomized pa-
mencing intervention as soon as patients are medically
tients, rather than the tube itself (Leder et al. 2005,
viable. This new approach has the potential ultimately
Sharma et al. 2007). This new understanding has been
to help reduce decannulation time, as secretion control
derived from a number of studies investigating the im-
and airway protection often remain as factors limiting
pact of changing respiratory and tracheostomy condi-
the progression to safe removal of the tracheostomy tube
tions on swallowing. Whilst methodological differences
long after the physiological (respiratory) need for the
(i.e., underlying aetiologies, clinical populations, sample
tube has resolved. Indeed, research has confirmed that
sizes, tracheostomy conditions, research design) limit
the presence of dysphagia is associated with increased
firm conclusions, evidence suggests that while swal-
time to decannulation (Romero et al. 2010).
lowing biomechanics may be altered by changing tra-
While it is recognized anecdotally that clinical
cheostomy conditions, there is no change in function.
practice patterns have changed, there is limited in-
For example, while increasing cuff pressures have been
formation regarding the patterns of clinical manage-
shown to impact swallowing reflexes, subsequent im-
ment of dysphagia for tracheostomized patients. Despite
pact on airway protection was not examined or reported
558
recognition that aspiration is variable but common in
this population (30–65%) (Ding and Logemann 2005,
Leder 2002, Schonhofer et al. 1999), there are minimal
reported data to benchmark services or guide prognostic
and management decisions.
The few studies to date have concentrated on re-
porting whole cohort data, with little recognition of
population demographics and how that may impact the
data. A retrospective audit of 140 patients from mul-
tiple diagnostic groups (including neurological, head
and neck, cardiothoracic and respiratory) who had re-
ceived a tracheostomy acutely reported that oral in-
take occurred at day 15, and decannulation at day
16 post-tracheostomy (Freeman-Sanderson et al. 2011).
However, the study largely focused on the patterns of
performance for the total heterogeneous group, and
whilst the timing of milestones differed for clinical pop-
ulations these differences were not explored. More re-
cently, Sutt et al. (2015) retrospectively examined the
impact of inline speaking valves on the duration of ven-
tilation and commencement of verbal communication
in a cardiorespiratory cohort of 129 patients; oral in-
take return in this audit was closer to 1 week post-
tracheostomy.
The presence and severity of dysphagia in patients
without a tracheostomy varies significantly according to
underlying aetiology, therefore, it stands to reason that
patterns of oral intake (and decannulation) will also vary
across patients from different diagnostic groups with a
tracheostomy in situ. As such, systematic examination
of the return to oral intake for clinical populations who
undergo tracheostomy is required to explore this issue
further; in doing so providing baseline information to
guide practice patterns and evaluate new interventions.
Therefore, the aim of the study was to describe clin-
ical management patterns for a heterogeneous cohort of
patients who received a new tracheostomy in an acute in-
patient setting, and to examine the nature of variability
across specialty clinical populations. Specific aims were:
(1) to describe patient demographics; (2) to describe the
attainment and timing of milestones related to nutrition,
cannulation and length of stay (LOS); (3) to describe
clinical management patterns related to the commence-
ment of oral intake and cessation of EN throughout
hospital admission and along the decannulation path-
way; and (4) to identify relationships between variables.
Methods
The study was conducted at the Royal Adelaide Hospi-
tal (RAH), a 640-bed acute tertiary hospital that admits
adult inpatients under multiple specialist caseloads (544
general medical and surgical beds, 96 specialty beds in-
cluding head and neck surgery, neurosurgery, cardio-
thoracic surgery, and state-wide services for spinal cord
Lee Pryor et al.
injury (SCI) and burns). A 12-month retrospective re-
view of clinical management patterns relating to oral
feeding and decannulation was conducted on patients
who received a new tracheostomy between January and
December 2009. During this time there were more than
34 000 overnight hospital admissions, including 3217
admissions to the intensive care unit (ICU), which com-
prised 32 ICU beds (24 general, eight cardiothoracic)
and 10 high-dependency beds.
In this service, patients with a tracheostomy will
remain in ICU until they are weaned from mechani-
cal ventilation; they will then be transferred to one of
several tracheostomy cohort wards for the remainder of
their hospital admission. Exceptions to this are patients
who require long-term ventilation (e.g., SCI); and head
and neck surgical patients who may go via ICU or high
dependency for postoperative monitoring, or straight
to the ward from the recovery suite. Initial deflation of
the tracheostomy cuff predominantly occurs once me-
chanical ventilation is no longer required and when oral
secretions are overtly tolerated. Cuff deflation is con-
tinuous at first attempt unless poorly tolerated, as such;
periodic cuff deflation with re-inflation for ventilation is
not common practice in this setting. Once on the ward,
patients who require ongoing therapy will be subse-
quently discharged to a rehabilitation facility, preferably
following decannulation.
SLT involvement with tracheostomized patients is
via medical referral while in ICU, and blanket referral
on the ward. The SLT is a member of the intensivist-led
multidisciplinary tracheostomy team which reviews all
patients whose tracheostomy remains in situ following
discharge from ICU, excluding those under the care of
the ears, nose and throat (ENT) surgeons. Assessment
by an SLT will occur for any patient who is deemed med-
ically suitable for oral intake regardless of tracheostomy
cuff status, through bedside assessment ± instrumental
FEES as needed. Of relevance to this study is the deliber-
ate minimal use of thickened fluids with patients in this
cohort. SLTs in this service follow the principles of the
Frazier Free Water Protocol (Panther 2005) and as such
it is common practice for aspirating or ‘at risk’ patients
to be allowed free water (and/or ice chips) provided oral
hygiene is optimized and there is no obvious patient dis-
tress or negative sequelae. Therefore, oral intake in this
paper is referred to as any diet/fluid other than water,
as the latter is routinely offered to patients who have a
tracheostomy unless medically contraindicated.
Participants
Ethical approval was obtained from the RAH Hu-
man Research Ethics Committee to access informa-
tion recorded within the medical and SLT records
(progress notes, operation records, ICU charts, ICU
Oral intake and decannulation post-tracheostomy
Paent idenficaon:
DRG A06
41880-00 percutaneous tracheostomy
41881-00 open tracheostomy, temporary
41881-01 open tracheostomy, permanent
n=211
Incorrectly coded n=19
Paent idenficaon:
Tracheostomy in-situ n=192
Study cohort n=126
* same patients incorrectly coded in DRG
# majority deceased with tracheostomy in-situ and prior to oral intake
Figure 1. Study cohort.
medical transfer summaries, ward tracheostomy obser-
vation charts), and the study carried out in accordance
with the Declaration of Helsinki. All patients catego-
rized by RAH case mix coding as diagnostic related
group (DRG) A06 were cross-referenced with nursing
records to identify patients recorded as having a tra-
cheostomy in situ (n = 192). Following the exclusions
detailed in figure 1 a cohort of 126 patients remained
for analysis.
Data collection
Demographic information was collected alongside data
relating to the timing of milestones and patterns of clin-
ical management, as detailed in table 1. Data were col-
lected and entered into a Microsoft Excel spreadsheet
by an independent research assistant who had experi-
ence conducting clinical chart audits, with 25% of the
data subsequently cross-checked by the lead investigator.
Data error was determined to be 3%, which was deemed
acceptable. Analysis of all variables for the cohort was
conducted using descriptive statistics, specifically re-
porting mean with standard deviation (SD), or median
and range for temporal data; and frequency counts (ex-
pressed as percentages) for data reporting patterns of
practice. Relationships between temporal variables were
559
Paent idenficaon:
Hospital nursing records
n=205
Incorrectly coded n=13*
EXCLUSIONS
Tracheostomy in-situ on
admission n=25
Total laryngectomy n=12
Mini-tracheostomy n=1
Casenotes unavailable n=5
Deceased n=25#
calculated using the Pearson correlation coefficient with
the Statistical Package for the Social Sciences (SPSS,
v.21.0), where r-values > .6 were deemed a substantial
correlation (Levin and Fox 2006). Subsequent descrip-
tive analysis by clinical population was also undertaken
across the collected variables.
Results
Cohort demographics
Table 2 contains the demographics of the cohort. Me-
dian age was 53 (18–87) years, with mean APACHE
III score 60 (SD = 27) (APACHE = Acute Physiology,
Age, Chronic Health Evaluation). Within the cohort,
patients were sub-classified into the eight clinical pop-
ulations representing the specialty areas of the RAH
service: (1) head and neck surgery, (2) neurosurgery, (3)
general medicine, (4) general surgery, (5) cardiothoracic
surgery, (6) spinal cord injury, (7) neurology, and (8)
burns. The two most commonly occurring clinical pop-
ulations were head and neck surgery, and neurosurgery
(table 2). Most patients (90%) were managed with a
tracheostomy tube that had an above-cuff suctioning
port as both their initial and replacement tube; with
a minority requiring an extended length tube (10%)
560 Lee Pryor et al.
Table 1. Data collection parameters

Clinical data Definition

Demographics
Acute Physiology, Age, Chronic Health Severity of disease classification score at the time of intensive care unit
Evaluation (APACHE III) (ICU) admission
Method of tracheostomy insertion Percutaneous or surgical
Tracheostomy tube type Shiley Evac (where normal anatomy expected) or Shiley XLT (extended
length tube for anatomical variation)
Tracheostomy tube size Manufacturer size 8 or 6
Timing of milestones
ICU length of stay ICU admission to ICU discharge (days)
Hospital length of stay Hospital admission to hospital discharge (days)
Length of intubation Insertion of endotracheal tube to tracheostomy cannulation (days)
Length of tracheostomy cannulation Tracheostomy insertion to removal (days)
Time to tracheostomy cuff deflation Tracheostomy insertion to continuous cuff deflation (days)
Time to initial oral intake Tracheostomy insertion to commencement of oral fluids and/or diet
(days)—excluding water only
Time to enteral nutrition (EN) cessation Tracheostomy insertion to complete cessation of EN (days)
Clinical management patterns
Decannulation achieved by discharge Percentage decannulated by ICU and hospital discharge
Oral intake achieved by discharge Percentage commenced on oral intake by ICU and hospital discharge
EN cessation by discharge Percentage cessation EN by ICU and hospital discharge
Tracheostomy status at initial oral intake and/or (1) Cuff inflated, (2) cuff deflated, (3) uncuffed ± occluded, (4)
EN cessation post-decannulation
Decannulation pathway followed Percentage cuff deflation only; percentage downsizing ± occlusion

Table 2. Cohort demographics (n = 126) tion. Considerable individual variability was evident, yet
Demographics n (%) Median (range) Mean ± SD
distinct patterns relating to LOS, intubation durations,
and timing of tracheostomy and nutrition milestones
Gender
Male 82 (65)
were evident across clinical populations (table 3).
Female 44 (35) For most clinical populations, LOS in the ICU and
Clinical population Age (years) APACHE III hospital averaged 2–3 weeks and 1–2 months respec-
Head and neck surgery 41 (33) 58 (19–87) 64 ± 23 tively. Exceptions to this pattern were the head and neck
Neurosurgery 29 (23) 45 (18–71) 62 ± 23 surgery population who stayed in ICU for fewer than
General medicine 18 (14) 65 (24–73) 85 ± 30 1–2 days (if at all) and were discharged after 2 weeks,
General surgery 16 (13) 56 (21–73) 64 ± 23 and the patients with burn injury who had protracted
Cardiothoracic surgery 8 (6) 63 (51–79) 60 ± 12
Spinal cord injury 6 (5) 33 (18–45) 41 ± 12 ICU admissions of over 7 weeks and hospital LOS of 3–
Neurology 5 (4) 58 (48–75) 52 ± 9 4 months or more. These two clinical populations also
Burns 3 (2) 47 (32–58) 74 ± 31 had the shortest and longest durations to commenc-
Insertion methoda ing cuff deflation and final decannulation respectively.
Percutaneous 54 (45) General surgical and SCI patients were the earliest to
Surgical 69 (55) commence oral intake within 7 days post-tracheostomy;
Tube type
Shiley Evac 113 (90)
patients with burns and neurological diagnoses took the
Shiley XLT 13 (10) longest (> 3 weeks post-tracheostomy). Prolonged pe-
Tube size riods of dual oral nutrition and EN were observed in
8 97 (77) clinical populations who had high nutritional require-
6 29 (33) ments post-injury; patients with SCI took 9 days to
Notes: a Insertion method was unknown for three patients. transition to oral intake and two of the three burns pa-
tients still required dual modality nutrition on hospital
due to anatomical variation. Nearly all patients were as- discharge.
sessed by SLT during admission (87%, n = 109), with
the tracheostomy still in situ for the majority of initial
assessments (90%, n = 98). Clinical management patterns
Overall, 43% of the cohort started oral intake while in
Timing of milestones
ICU (n = 54) (table 4). By hospital discharge 86% were
Table 3 details the timing of milestones for the cohort in eating and drinking with 14% (n = 18) never resuming
relation to hospital admission, cannulation and nutri- oral intake. Oral intake commenced in the ICU more
Oral intake and decannulation post-tracheostomy 561
Table 3. Timing of milestones for the total cohort and by clinical population (days, median and range)

Days post-tracheostomy insertion

ICU LOS Hospital LOS Intubation Cuff deflation Decannulation Oral intake EN cessation
Clinical population Median (range) Median (range)
a b
Head and neck surgery 1.7 18.5 0 4 12 10 12
(0.5–39) (7–103) (0–8) (0–25) (7–25) (1–25) (4–27)
Neurosurgery 13.8 53 8 8.3 18 13.5 16
(6–21) (22–150) (1–20) (2–53) (5–66) (5–39) (5–49)
General medicine 23 49.5 11 7.5 13 10 14
(3.5–65) (23–150) (1–25) (3–32) (6–52) (5–24) (5–34)
General surgery 19.4 35 7 12 15 7 13
(4–31) (17–78) (1–15) (3–20) (6–23) (3–13) (5–23)
Cardiothoracic surgery 25 55 14 17 19 18 19
(16–50) (45–113) (9–24) (6–67) (9–88) (7–94) (14–41)
Spinal cord injury 28.7 65 10 10.5 20 6 15
(16–35) (37–82) (4–17) (4–12) (11–63) (4–32) (13–63)
Neurology 15.1 37 8 12 19 21.5 25
(9–52) (28–113) (1–14) (2–17) (11–89) (15–59) (15–28)
Burns 51.5 126 24 22 33 29 n.a.d
(34–54) (77–196) (9–26) (18–35) (23–67) (22–37)
Total cohort n = 112a n = 126 n = 89b n = 124 n = 108 n = 108c n = 89e
14.7 38 9 7.5 15 10.5 17
(0.5–65) (7–196) (1–26) (0–67) (5–89) (1–75) (4–63)
Notes: EN, enteral nutrition.
a
Fourteen patients did not go via ICU.
b
Thirty-seven patients were intubated at the time of surgery in which the tracheostomy performed.
c
Eighteen patients did not return to oral intake.
d
Two patients were discharged on dual oral and EN.
e
A total of 121 patients received EN during admission; five did not require EN.

Table 4. Proportion of patients who achieved oral intake and decannulation at the time of (a) ICU discharge and (b) hospital
discharge, reported for both the total cohort and each clinical population

Discharge from ICU Discharge from hospital

Clinical Oral intake, Decannulation, Oral intake, Decannulation,
population n n (%) n (%) n (%) n (%)
Head and neck surgery 41 5a (12) 1a (4) 37 (90) 29 (73)
Neurosurgery 29 9 (31) 3 (10) 17 (59) 25 (86)
General medicine 18 12 (67) 11 (61) 17 (94) 17 (94)
General surgery 16 15 (94) 11 (69) 16 (100) 16 (100)
Cardiothoracic surgery 8 6 (75) 7 (88) 8 (100) 8 (100)
Spinal cord injury 6 5 (83) 2 (33) 6 (100) 5 (83)
Neurology 5 1 (20) 0 (0) 4 (80) 5 (100)
Burns 3 1 (33) 1 (33) 3 (100) 3 (100)
Total cohort 126 54 (43) 36 (29) 108b (86) 108 (86)
Notes: a Fourteen patients did not go via ICU.
b
Eighteen patients did not return to oral intake.

often for patients managed by the general medical, gen-
eral surgical, cardiothoracic and SCI teams. In contrast,
head and neck, neurosurgical and neurological popula-
tions were more likely to commence oral intake when
on the ward. Neurosurgical patients were least likely
to have recommenced oral intake by hospital discharge
(table 4).
Figure 2 displays the commencement of oral intake
and cessation of EN relative to tracheostomy status for
the total cohort. For those who resumed oral intake (n =
108) the majority (82%) commenced oral intake with
a tracheostomy in situ, though typically in either a cuff
deflated or uncuffed/cork condition. Most clinical pop-
ulations followed this pattern, however variations were
observed for the general surgical and SCI populations
where greater proportions (44% and 83% respectively)
commenced oral intake with the cuff inflated (table 5).
Almost all patients (96%, n = 121) received some
nutritional support via enteral feeding during their
hospital admission; with 74% (n = 89) transition-
ing successfully to oral intake. Of this group, EN was
ceased for more than half (56%) after decannulation
562 Lee Pryor et al.

Figure 2. Proportion of patients commencing oral intake and ceasing enteral nutrition at various stages along the decannulation pathway.
‘decan’ represents the stage post-decannulation.

Table 5. Oral intake commencement by tracheostomy status and clinical population (n = 108)

Clinical Cuff inflated Cuff deflated Uncuff ± cork Decannulated

population (n = 15) (n = 36) (n = 38) (n = 19)
Head and neck surgery 3 9 16 9
Neurosurgery – 6 7 4
General medicine – 9 7 1
General surgery 7 6 3 –
Cardiothoracic surgery – 4 1 3
Spinal cord injury 5 – 1 –
Neurology – – 3 1
Burns – 2 – 1

(figure 2). Despite 86% of the cohort commencing oral
intake, 25% (n = 32) were discharged with a percu-
taneous endoscopic gastrostomy (PEG) in situ. Those
who remained PEG dependent at discharge consisted
primarily of two clinical populations: with almost half
comprising neurosurgical patients (n = 15) and a third
post-head and neck surgery (n = 11).
Decannulation
A total of 14% of patients in the total cohort were dis-
charged from hospital with the tracheostomy still in situ
(n = 18). Of the 108 who achieved decannulation, a
third were decannulated in ICU (33%, n = 36) and
were primarily from cardiothoracic, general surgical and
general medical populations (table 5). The majority of
patients were decannulated on the ward (67%, n =
72). Most were downsized to a smaller uncuffed tra-
cheostomy prior to tube removal (76%, n = 82), and
of these, nearly all underwent tube occlusion (i.e., cork-
ing) as part of the decannulation process (89%, n = 73).
Only a small proportion of the cohort (24%, n = 26)
were decannulated following a period of cuff deflation
only. The pattern of downsizing and occlusion prior
to decannulation was utilized as the dominant path-
way in most clinical populations with the exception of
cardiothoracic and general surgical patients (figure 3).
However, it should be noted that 85% of the patients
decannulated via the cuff deflation only pathway were
decannulated while in ICU (n = 22). That is, 61% of
patients decannulated in ICU did so after cuff deflation
only compared with 6% on the wards. The overall recan-
nulation rate was 4.6% (n = 5), of these, two patients
were recannulated in ICU and three on the ward. One
patient required a permanent tracheostomy.
Relationships between variables
Correlations were used to explore relationships between
variables. Using the full cohort data, time to commence-
ment of oral intake had a strong correlation (r = .805,
p = .001) with time to decannulation, which in turn
Oral intake and decannulation post-tracheostomy
% cuff deflation only
100 7
16
24
80
60
% 93
40 84
76
20
0
Clinical populations
Figure 3. Proportion of patients in each clinical population managed by either of the two decannulation pathways.
was associated with increased length of hospital stay (r =
.687, p = .006). Small numbers prevent correlations to
be computed for all clinical populations, however of the
four largest subgroups, three had substantial correlations
between time to oral intake and time to decannulation
(neurosurgery r = .879 p = .000; general medicine
r = .611, p = .016; head and neck surgery r = .603,
p = .001). No significant relationship (r = .16, p =
.553) between these variables was observed in the general
surgery group, in keeping with the early introduction of
diet/fluids while the cuff remained inflated.
Discussion
Analysis revealed differential clinical patterns between
patient populations for oral intake attainment and de-
cannulation, supporting the hypothesis that distinct pat-
terns of management can be anticipated across different
clinical populations; each with differing airway require-
ments and concerns. Overall, the vast majority of pa-
tients who received a tracheostomy acutely were able
to recommence oral intake during hospital admission,
most while the tracheostomy was still in situ, and the
majority achieved decannulation prior to hospital dis-
charge. Time to decannulation was comparable to other
centres with heterogeneous groups (Choate et al. 2009,
Freeman-Sanderson et al. 2011); however, time to com-
mencement of oral intake was earlier than prior research
(Freeman-Sanderson et al. 2011) and longer than con-
temporary work (Sutt et al. 2015), likely reflecting the
563
% downsize +/- occlusion
40
56
86
100 100
60
44
14
variation in cohort demographics between the studies.
Oral intake was introduced at every stage of the tra-
cheostomy pathway, including prior to cuff deflation.
Timing of oral intake commencement was evenly split
between intensive care and the ward and seemed to be
influenced by the clinical profile of the population, not
location.
The timing of oral intake commencement relative
to the decannulation pathway appeared to be largely re-
lated to the tracheostomy weaning and/or decannulation
goals, which in turn related to the underlying diagnosis.
The typical decision-making processes of the treating
SLTs in the setting where this study was conducted are
visually represented in figure 4. That is, if cuff deflation
were imminent and/or dysphagia likely; cuff deflation
occurred first for evaluation of secretion management
followed by assessment for commencement of oral nu-
trition. Clinically this is the period where restored upper
airway flow enables greater functionality (speech, swal-
lowing, sensation, olfaction, taste, coughing) but also
may improve the ease of assessment and identification
of factors that could influence subsequent progression,
such as vocal cord paresis (Hales et al. 2008). It is not
unexpected, therefore, that the majority of patients in
the cohort who were able to commence oral intake, did
so following cuff deflation or with an uncuffed tube in
situ. This was the primary pattern for all clinical pop-
ulations with the exception of SCI. With populations
for whom ventilator support was ongoing, cuff defla-
tion not imminent, and/or dysphagia not overt—such
564 Lee Pryor et al.

 Medical consent for oral intake

 No contraindicaons for oral intake
 Paent sufficiently alert

CUFF INFLATED CUFF DEFLATED / UNCUFFED

Is cuff deflaon Is paent tolerang

imminent? secreons?

NO YES NO YES

Is paent tolerang
Assess for oral intake
secreons?

YES NO Connue NBM

Assess for oral intake Assess for oral intake

with cuff inflated post cuff deflaon

Figure 4. Clinical decision-making process regarding timing of oral intake commencement.

as SCI and a large proportion of general surgical pa-
tients (44%)—oral intake was commenced whilst the
tracheostomy cuff remained inflated for nutritive and
comfort intent (figure 4).
The current analysis also confirmed an association
between time to oral intake and time to decannulation.
Whilst duration to commencing oral intake is not a di-
rect measure of dysphagia per se, and is acknowledged
to be influenced by other factors, the data is in keeping
with previous research linking the presence of dyspha-
gia to increased cannulation time (Romero et al. 2010).
Furthermore, prolonged time to decannulation corre-
lated with increased length of hospital stay, which may
reflect discharge barriers to rehabilitation and residen-
tial care posed by tracheostomy and/or nasogastric tube
presence; with over half those able to wean from EN not
doing so until after tracheostomy removal, and a quar-
ter of patients discharged on long-term nutrition via a
gastrostomy.
Of those who did not return to oral intake by hospi-
tal discharge, the majority had a severe brain injury with
reduced levels of consciousness the presumed barrier to
oral feeding. A small number of patients post-head and
neck surgery were discharged nil by mouth (NBM) yet
all were on sips of water for comfort; as such, while they
didn’t return to oral intake from a nutrition point of
view they weren’t completely NBM either. The remain-
ing two patients had severe bulbar dysfunction as a result
of neurological conditions and were unable to return to
any oral intake (including water); one of these patients
was admitted under the general medicine bed-card due
to the progressive nature of the neurological condition.
Despite clear patterns of management within the
cohort, further examination revealed differential pat-
terns between clinical populations related to where and
when oral intake was commenced, and at what stage
in the tracheostomy pathway. Patients who sustained
acute trauma and required a cuffed tracheostomy to fa-
cilitate prolonged ventilation or airway clearance (i.e.,
SCI and/or general surgery) nearly all started oral intake
within a week of tracheostomy insertion, while in ICU,
and as previously stated—many with the tracheostomy
cuff inflated. These patients were typically alert and cog-
nitively intact. The incidence of dysphagia with acute
tetraplegia post-SCI has been previously reported as vari-
able but common (16–41%) (Seidl et al. 2010, Shem
et al. 2011); only one patient in the study was unable to
commence oral intake early, with dysphagia related to a
neuropraxic injury to the recurrent laryngeal nerve sus-
tained during anterior cervical surgery. Although a small
cohort, the data is consistent with the 16% incidence of
dysphagia previously reported (Seidl et al. 2010). Dual
oral nutrition and EN was required for a prolonged dura-
tion despite early oral intake, reflecting high nutritional
requirements in the SCI population (Wong et al. 2012).
Furthermore, compounding factors such as dependence
on others for oral feeding, and emotional influences in-
cluding demotivation in the context of coming to terms
with diagnosis and longer term prognosis, could not be
discounted.
Oral intake and decannulation post-tracheostomy
Allowing patients to eat and drink with an inflated
cuff may be controversial in some services. However, in
the absence of medical contraindications or robust ev-
idence, patients who are functionally capable are given
the opportunity to consume oral intake for comfort and
quality of life purposes—and nutrition—in the partic-
ipating ICU. A small subset of head and neck patients
requiring an emergency tracheostomy due to airway ob-
struction at the laryngeal level were also fed within a day
of cuffed tracheostomy insertion, also in ICU. These
patients were not dysphagic prior to surgery and the
presence of the tracheostomy tube did not alter func-
tion. As such, placement of a nasoenteric feeding tube
was avoided. Ready access to bedside FEES can assist de-
cision making in these instances (McGowan et al. 2007).
SLTs in the participating hospital can be accredited to
perform independent FEES, allowing direct visualiza-
tion of airway safety, and facilitating team discussion
and decision making regarding risks versus benefits of
commencing oral intake with an inflated cuff.
Patients under general medical and cardiothoracic
care predominantly commenced oral intake while in
ICU and required a tracheostomy for prolonged ven-
tilation, similar to the SCI and general surgical popu-
lations. However, these patients had longer periods of
intubation than all other diagnostic groups (excepting
burns), were the oldest populations within the cohort,
and in the case of the general medical patients had the
highest APACHE scores reflecting the highest severity of
illness at the point of admission to the ICU (table 2). We
know that prolonged intubation in the critically ill pop-
ulation is associated with a high incidence of dysphagia
post-extubation (41%) (Bordon et al. 2011) with in-
creasing age and number of ventilation days escalating
risk (Bordon et al. 2011). Therefore, an older patient
demonstrating failure to wean may already be predis-
posed to dysphagia, with recovery further complicated
post-tracheostomy by deconditioning, decompensation
with altered swallowing/respiratory patterns and comor-
bidities in keeping with an aged population with di-
minished functional reserve (Leder 2002, Martin-Harris
et al. 2005). These patients demonstrated good func-
tional recovery whilst in intensive care yet their ICU
LOS was longer than most (> 3 weeks) in keeping with
their medical fragility. Oral intake occurred primarily
following cuff deflation or with an uncuffed tube in situ,
and for a third of cardiothoracic patients after decannu-
lation. Oral intake timing for cardiothoracic patients in
the study was 18 days post-tracheostomy which is sim-
ilar to earlier research (20 days) (Freeman-Sanderson
et al. 2011) yet considerably longer than a recent study
describing commencement of oral intake within a week
of tracheostomy insertion (Sutt et al. 2015).
The lowest attainment of oral intake was noted in
the neurosurgical population, where the tracheostomy
565
was inserted for gross airway protection and respiratory
control following acquired brain injury in which re-
duced conscious state was a compounding factor. While
there was an even split between those able to start eat-
ing and drinking in the ICU following cuff deflation,
and those who transitioned to oral intake on the ward
as their alertness improved; a large number (41%) did
not return to oral nutrition at all. As such, the median
return to oral intake for this population of two weeks
post-tracheostomy is perhaps not a true reflection of
oral intake patterns of the overall neurosurgical popula-
tion. For those who did progress well there were large
gaps between milestones, i.e., multiple days separated
cuff deflation, oral intake, decannulation and cessation
of EN, reflecting slower patient recovery with cognitive
overlay.
The majority of patients with head and neck, burns
and neurological diagnoses commenced oral intake on
the ward, predominantly with the cuff deflated or with
an uncuffed tube in situ (and a quarter following de-
cannulation); yet underlying aetiology influenced the
considerable variation in attainment of oral intake and
decannulation milestones. Patients with neurological di-
agnoses primarily required the tracheostomy for airway
protection; as such oral intake was not considered un-
til secretion management was established and the pa-
tients were advanced in their decannulation pathway.
Oral intake was commenced three weeks following tra-
cheostomy, in keeping with slower incremental swallow-
ing recovery.
Of all clinical populations the head and neck surgi-
cal patients had the shortest ICU and hospital LOS,
time to cuff deflation, and decannulation; and, the
shortest time to oral intake commencement excepting
the SCI and general surgical populations previously de-
scribed. The tracheostomy was primarily placed as part
of a planned surgical procedure involving extensive re-
construction of upper airway structures. Although cuff
deflation occurred early, active swallowing of diet or
fluids was avoided for a medically dictated period of
time (usually 10 days) to ensure integrity of any surgical
restoration (e.g., flap) and resolution of postoperative
oedema. As a result, oral intake tended to be intro-
duced closer to decannulation. Some degree of swallow-
ing morbidity was expected yet nearly all (90%) were dis-
charged home on some oral intake; furthermore these
patients were quick to transition from EN. This may
be in part due to the planned nature of their admis-
sion and surgery, with known expectations of swallow-
ing changes prior to surgery. In addition, patients were
typically mobile and had a strong protective cough, and
therefore may have had a high threshold for tolerance
of aspiration. Despite having the lowest success of de-
cannulation of all clinical populations (due to ongo-
ing airway obstruction or as a precaution for adjuvant
566
chemoradiotherapy); the timing of decannulation for
those that could have the tube removed, was the quick-
est of all subgroups, consistent with previous research
(Leung et al. 2003).
The longest milestone durations without exception
were observed in patients with burns and associated
inhalation injuries, in keeping with earlier research
(Rumbach et al. 2012). In this population a tra-
cheostomy may be required not only for prolonged
ventilation but also for airway patency, due to laryn-
gotracheal pathology such as laryngeal oedema and mu-
cosal changes resulting from inhalation burn (Clayton
et al. 2010). These issues can in turn impact swallowing
recovery (Clayton et al. 2010; Rumbach et al. 2012).
However, for this group, medical factors may influence
the observed temporal recovery patterns, including the
need to maintain a patent airway throughout multiple
surgical debridement procedures. Consistent with prior
literature (Rumbach et al. 2012), this patient cohort
was also observed to have prolonged durations of EN.
EN is necessary due to the high energy requirements
post-burn injury, and its prolonged use is largely unre-
lated to swallowing status. In fact, despite returning to
oral intake, two of the three patients were discharged to
rehabilitation with dual oral nutrition and EN.
The decannulation pathway followed by the overall
cohort favoured downsizing and occlusion. The popula-
tions decannulated after a period of cuff deflation only,
tended to be those for whom the tracheostomy was
inserted for prolonged ventilation, and not airway pro-
tection or patency (i.e., cardiothoracic, general surgery,
SCI). However pathway differences may also be influ-
enced by individual preferences of the clinical staff mak-
ing the decisions, as well as the location of care; with
nearly all patients decannulated after cuff deflation only
noted to reside in ICU at the time of tube removal. This
may reflect differences in environment, resources, staff–
patient ratios and skill mix between ICU and the ward;
with immediate access to staff with specialist skills and
experience in emergency airway management in ICU
potentially influencing decision-making.
To the authors’ knowledge this is the first paper
specifically to present the differential patterns of return
to oral intake in a cohort of acutely tracheostomized
patients from multiple clinical populations. Oral intake
was commenced for each patient when they were indi-
vidually ready to start eating and drinking, regardless of
tracheostomy cuff status, based on clinical assessment
and risk analysis of the swallowing function and poten-
tial confounding factors of each individual. It should be
noted that cuff deflation practices of the participating
hospital may have had some bearing not only on the tim-
ing of cuff deflation, but also on the decision-making
around timing of oral intake, as the cuff typically re-
mains inflated until the patient has been liberated from
Lee Pryor et al.
ventilation. Further prospective investigation of the im-
pact of cuff inflation status on swallowing function is
needed, however the observational data of clinical prac-
tice would suggest that there is a role for oral intake in
some patients with a cuffed tracheostomy in situ.
Whilst the paper provides useful data for generation
of power calculations for future research, we acknowl-
edge the study is limited by small numbers, particularly
in some patient populations, and its retrospective de-
sign. Additional data on levels of consciousness (e.g.,
severity, duration) would have aided the interpretation
of neurosurgical data and should be captured in future
studies; similarly data on surgical intervention post-
tracheostomy for the burns population (e.g., number
and frequency of returns to theatre) should be recorded.
The timing of milestones may have been influenced not
only by the clinical population and underlying diagno-
sis but also by the individual management preferences
of the treating clinician(s) and the timing of referral,
the latter of which has not been investigated nor ad-
dressed in this paper. The practice patterns described,
however, are consistent with clinical observations over
ten years of service provision to these specific patient
populations.
Conclusions
While the data provide group patterns for an overall co-
hort, considerable variability in the attainment of mile-
stones related to cuff deflation, decannulation, oral in-
take and EN were evident between and within clinical
populations. The data would suggest that tracheostomy
patients should not be looked upon as a singular cohort;
rather, evaluation of factors with specific attention made
to underlying aetiology and individual clinical presen-
tation is essential. The role of the SLT is suited to this
aspect of a patient’s care within an acute setting and
multidisciplinary team.
Acknowledgements
Declaration of interest: The authors report no conflicts of interest.
The authors alone are responsible for the content and writing of the
paper.
References
ADHIKARI, N. K. and RUBENFELD, G. D., 2011, Worldwide demand
for critical care [Review]. Current Opinions in Critical Care,
17, 620–625. doi:10.1097/MCC.0b013e32834cd39c
AMATHIEU, R., SAUVAT, S., REYNAUD, P., SLAVOV, V., LUIS, D.,
DINCA, A., TUAL, L., BLOC, S. and DHONNEUR, G., 2012,
Influence of the cuff pressure on the swallowing reflex in tra-
cheostomized intensive care unit patients. British Journal of
Anaesthesia, 109, 578–583.
BONANNO, P. C., 1971, Swallowing dysfunction after tracheotomy.
Annals of Surgery, 174, 29–33.
Oral intake and decannulation post-tracheostomy
BORDON, A., BOKHARI, R., SPERRY, J., TESTA, D., FEINSTEIN, A.
and GHAEMMAGHAMI, V., 2011, Swallowing dysfunction af-
ter prolonged intubation: analysis of risk factors in trauma
patients. American Journal of Surgery, 202, 679–683.
BRUMMEL, N. E., BALAS M.C., MORANDI, A., FERRANTE, L., GILL
T.M. and ELY, E. W., 2015, Understanding and reducing
disability in older adults following critical illness. Critical
Care Medicine, 43, 1265–1275.
CHOATE, K., BARBETTI, J. and CURREY, J., 2009, Tracheostomy de-
cannulation failure rate following critical illness: a prospective
descriptive study. Australian Critical Care, 22, 8–15.
CLAYTON, N., KENNEDY, P. and MAITZ, P., 2010, The severe burns
patient with tracheostomy: implications for management of
dysphagia, dysphonia and laryngotracheal pathology. Burns,
36, 850–855.
DAGGETT, A., LOGEMANN, J., RADEMAKER, A. and PAULOSKI, B.,
2006, Laryngeal penetration during deglutition in normal
subjects of various ages. Dysphagia, 21, 270–274.
DAVIS, D. G., BEARS, S., BARONE, J. E., CORVO, P. R. and
TUCKER, J. B., 2002, Swallowing with a tracheostomy tube
in place: does cuff inflation matter? Journal of Intensive Care
Medicine, 17, 132–135.
DING, R. and LOGEMANN, J., 2005, Swallow physiology in patients
with trach cuff inflated or deflated: a retrospective study. Head
and Neck, 27, 809–813. doi:10.1002/hed.20248
FREEMAN, B. D., ISABELLA, K., LIN, N. and BUCHMAN, T. G., 2000,
A meta-analysis of prospective trials comparing percutaneous
and surgical tracheostomy in critically ill patients. Chest, 118,
1412–1418.
FREEMAN-SANDERSON, A., TOGHER, L., PHIPPS, P. and ELKINS,
M., 2011, A clinical audit of the management of patients
with a tracheostomy in an Australian tertiary hospital inten-
sive care unit: focus on speech–language pathology. Interna-
tional Journal of Speech–Language Pathology, 13, 518–525.
doi:10.3109/17549507.2011.582520
HALES, P., DRINNAN, M. and WILSON, J., 2008, The added value
of fibreoptic endoscopic evaluation of swallowing in tra-
cheostomy weaning. Clinical Otolaryngology, 33, 319–324.
LEDER, S., 2002, Incidence and type of aspiration in acute care pa-
tients requiring mechanical ventilation via a new tracheotomy.
Chest, 122, 1721–1726. doi:10.1378/chest.122.5.1721
LEDER, S. B., JOE, J. K., ROSS, D. A., COELHO, D. H. and MENDES,
J., 2005, Presence of a tracheotomy tube and aspiration status
in early, postsurgical head and neck cancer patients. Head and
Neck, 27, 757–791.
LEUNG, R., CAMPBELL, D., MACGREGOR, L. and BERKOWITZ, R.,
2003, Decannulation and survival following tracheostomy
in an intensive care unit. Annals of Otology, Rhinology and
Laryngology, 112, 853–859.
LEVIN, J. and FOX, J. A., 2006, Correlation. Elementary Statistics
in Social Research, 10th edn (Boston, MA: Pearson, Allyn &
Bacon).
567
MARTIN-HARRIS, B., BRODSKY, M. B., MICHEL, Y., FORD, C. L.,
WALTERS, B. and HEFFNER, J., 2005, Breathing and swallow-
ing dynamics across the adult lifespan. Archives of Otolaryn-
gology and Head and Neck Surgery, 131, 762–770.
MCGOWAN, S. L., GLEESON, M., SMITH, M., HIRSCH, N. and
SHULDHAM, C. M., 2007, A pilot study of fibreoptic en-
doscopic evaluation of swallowing in patients with cuffed tra-
cheostomies in neurologic intensive care. Neurocritical Care,
6, 90–93.
PANTHER, K., 2005, The Frazier Free Water Protocol. Swallowing
and Swallowing Disorders, 14, 4–9.
ROBBINS, J., HAMILTON, J., LOF, G. and KEMPSTER, G., 1992,
Oropharyngeal swallowing in normal adults of different ages.
Gastroenterology, 103, 823–829.
ROMERO, C., MARAMBIO, A., LARRONDO, J., WALKER, K., LIRA, M.
T., TOBAR, E., CORNEJO, R. and RUIZ, M., 2010, Swallow-
ing dysfunction in nonneurologic critically ill patients who
require percutaneous endoscopic dilatational tracheostomy.
Chest, 137, 1278–1282. doi:10.1378/chest.09-2792
RUMBACH, A. F., WARD, E. C., CORNWELL, P. L., BASSETT, L. V. and
MULLER, M. J., 2012, Clinical progression and outcome of
dysphagia following thermal burn injury: a prospective cohort
study. Journal of Burn Care and Research, 33, 336–346.
SCHONHOFER, B., BARCHFELD, T., HAIDL, P. and KOHLER, D., 1999,
Scintigraphy for evaluating early aspiration after oral feeding
in patients receiving prolonged ventilation via tracheostomy.
Intensive Care Medicine, 25, 311–314.
SEIDL, R. O., NUSSER-MULLER-BUSCH, R., KURZWEILL, M. and
NIEDEGGEN, A., 2010, Dysphagia in acute tetraplegics: a ret-
rospective study. Spinal Cord, 48, 197–201.
SHARMA, O. P., OSWANSKI, M. F., SINGER, D., BUCKLEY, B., COUR-
TRIGHT, B., RAJ, S., WAITE, P. J., TATCHELL, T. and GANDAIO,
A., 2007, Swallowing disorders in trauma patients: impact of
tracheostomy. American Surgeon, 73, 1117–1121.
SHEM, K., CASTILLO, K., WONG, S. and CHANG, J., 2011, Dysphagia
in individuals with tetraplegia: incidence and risk factors.
Journal of Spinal Cord Medicine, 34, 85–92.
SUTT, A.-L., CORNWELL, P. C., MULLANY, D., KINNEALLY, T. and
FRASER, J. F., 2015, The use of tracheostomy speaking valves
in mechanically ventilated patients results in improved com-
munication and does not prolong ventilation time in cardio-
thoracic intensive care unit patients. Journal of Critical Care,
30, 491–494.
WARNECKE, T., SUNTRUP, S., TEISMANN, I. K., HAMACHER, C.,
OELENBERG, S. and DZIEWAS, R., 2013, Standardized endo-
scopic swallowing evaluation for tracheostomy decannulation
in critically ill neurologic patients. Critical Care Medicine, 41,
1728–1732.
WONG, S., DERRYL, F., JAMOUS, A., HIRANI, S. P., GRIMBLE, G. and
FORBES, A., 2012, The prevalence of malnutrition in spinal
cord injuries patients: a UK multicentre study. British Journal
of Nutrition, 108, 918–923.
Copyright of International Journal of Language & Communication Disorders is the property
of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to
a listserv without the copyright holder's express written permission. However, users may
print, download, or email articles for individual use.