European Archives of Oto-Rhino-Laryngology

https://doi.org/10.1007/s00405-017-4838-7

REVIEW ARTICLE

Timing of tracheostomy in patients with prolonged endotracheal

intubation: a systematic review
Ahmed Adly1 · Tamer Ali Youssef1 · Marwa M. El‑Begermy1 · Hussein M. Younis1

Received: 3 October 2017 / Accepted: 1 December 2017

© Springer-Verlag GmbH Germany, part of Springer Nature 2017

Abstract
The objective of this article is to evaluate the appropriate timing of tracheostomy in patients with prolonged intubationre-
garding the incidence of hospital-acquired pneumonia, mortality, length of stay in intensive care unit (ICU) and duration of
artificial ventilation. The study included published articles yielded by a search concerning timing of tracheostomy in adult
and pediatric patients with prolonged intubation. The search was limited to articles published in English language in the
last 30 years (between 1987 and 2017). For the 690 relevant articles, we applied our inclusion and exclusion criteria and
only 43 articles were included. 41 studies in the adult age group including 222,501 patients and 2 studies in pediatric age
group including 140 patients met our criteria. Studies in adult age group were divided into three groups according to the
methodology of determining the cut off timing for early tracheostomy, they were divided into studies that considered early
tracheostomy within the first 7, 14 or 21 days of endotracheal intubation, while in pediatric age group the cut off timing for
early tracheostomy was within the first 7 days of endotracheal intubation. There was a significant difference in favor of early
tracheostomy in adults’ three groups and pediatric age group as early tracheostomy was superior regarding reduced duration
of mechanical ventilation, with less mortality rates and less duration of stay in ICU. Regarding hospital-acquired pneumonia,
it was significantly less in adult groups but with no significant difference in pediatric age group (3 patients out of 72 pediatric
patient with early tracheostomy had pneumonia compared to 11 patients out of 68 with late tracheostomy). Studies defining
early tracheostomy as that done within 7 days of intubation had better results than those defining early tracheostomy as that
done within 14 or 21 days of intubation. In conclusion, early tracheostomy within 7 days of intubation should be done for
both adults and pediatric patients with prolonged intubation.

Keywords  Early tracheostomy · Tracheostomy timing · Prolonged endotracheal intubation · Tracheostomy

Introduction use of sedation and decreasing the duration of mechanical

Tracheostomy is a frequently performed procedure in about
24% of ICU patients as it has many advantages over pro-
longed endotracheal intubation as: reducing oropharyn-
geal and laryngeal trauma, reducing work of breathing by
decreasing airway dead space and lowering the airway resist-
ance, improving pulmonary secretion clearance, reducing the
Electronic supplementary material  The online version of this
article (https://doi.org/10.1007/s00405-017-4838-7) contains
supplementary material, which is available to authorized users.
* Tamer Ali Youssef
tayoussef@yahoo.com
1
Department of Otolaryngology, Ain Shams University, 36
Ismail Whaba Street, Naser City, District 9, Cairo, Egypt
ventilation and consequently length of ICU and hospital stay
[1, 2]. Open surgical tracheostomy and percutaneous trache-
otomy are methods of performing a tracheostomy in selected
patients. Open surgical tracheostomy may be required if per-
cutaneous tracheotomy technique is contraindicated due to
anatomic or other patient-related problems [2].
The optimal timing of tracheostomy in patients requiring
prolonged endotracheal intubation is still a debate despite its
advantages. The American National Association of Medical
Directors of Respiratory Care in 1989 stated that tracheostomy
is indicated if mechanical ventilation is to be continued for
more than 21 days while patients requiring mechanical ven-
tilation for less than 10 days are candidates for endotracheal
intubation [3]. Other authors recommended early tracheostomy
[4] while others mentioned that the evidence of the benefits of
tracheostomy over endotracheal intubation is insufficient [5].

13
Vol.:(0123456789)

The aim of this study is to conduct a meta-analytic study
to verify the appropriate timing of tracheostomy in patients
with prolonged intubation regarding the incidence of hospital-
acquired pneumonia, incidence of mortality, ICU length of
stay and duration of mechanical ventilation.
Methods
This study was conducted by searching medical literature in
the Medline database with the following keywords: timing of
tracheostomy, early versus late tracheostomy and tracheostomy
after endotracheal intubation.
The search was limited to articles published in the last
30 years (between 1987 and 2017). in English language to
identify prospective, randomized controlled studies assessing
the appropriate timing of tracheostomy in prolonged intubated
patients. Retrospective studies and studies without control
group were excluded. Different tracheostomy techniques either
open surgical or percutaneous tracheostomy are included in
the study.
Relevant articles were divided into adult and pediatric age
groups. Heterogeneity was found due to the difference in defin-
ing early and late tracheostomy. So, a sensitivity analysis was
conducted dividing the studies in adult age group into three
groups according to the methodology of determining the tim-
ing of early tracheostomy into studies that considered early
tracheostomy within the first 7 days of endotracheal intuba-
tion, studies that considered early tracheostomy within 14 days
of intubation and studies that considered early tracheostomy
within 21 days of intubation.
Each article included was analyzed regarding the methodol-
ogy and results (type of study, age of patient, number of cases
reported in the article, incidence of hospital-acquired pneumo-
nia, incidence of mortality, duration of mechanical ventilation
and length of ICU stay). The data collected from each article
were statistically analyzed utilizing ­MedCalc© version 15.8
(MedCalc© Software bvba, Ostend, Belgium).
Studies included in meta-analysis were tested for heteroge-
neity of the estimates using the following tests:
1. Cochran Q chi square test A statistically significant test
(p value < 0.1) denoted heterogeneity among the studies.
2. I-squared (I2) index which is calculated as follows:
( )
Q − df
I2 = × 100%
Q All favoring early tracheostomy over late tracheostomy
(Q Cochran Q chi square test result, df degree of freedom.
df = Number of studies − 1
3. The I-squared is interpreted as follows:
13
European Archives of Oto-Rhino-Laryngology
0–40%: unimportant heterogeneity; 30–60%: moderate
heterogeneity; 50–90%: substantial heterogeneity; 75–100%:
considerable heterogeneity.
Results
The literature search was performed on 12/2/2017 and
yielded 690 articles. Of these 690 articles, there were 647
articles irrelevant to our study from the title and abstract
and 43 articles were relevant by meeting our criteria. These
43 studies had 222,641 patients with 41 studies included in
adult age group with 222,501 patients and 2 studies included
in pediatric age group with 140 patients (Tables 1, 2).
Meta‑analysis of studies involving adult patients
Results of comparing early and late tracheostomy in adults
regarding hospital‑acquired pneumonia:
23 studies were involved including 154,048 patients were
divided into three groups according to the timing of early
tracheostomy.
I-squared (I2) index was 83% denoting considerable het-
erogeneity between studies.
Pooling of estimates via random effects method (REM)
showed that all studies regardless of the definition of early
and late tracheostomy showed an odds ratio (OR) of 0.68
(95% CI, CL = 0.6–0.77) which was statistically significant
(p value < 0.01) favoring early tracheostomy over late trache-
ostomy regarding incidence of hospital acquired pneumonia.
1. Studies defining early tracheostomy as that done within
7 days Pooling of estimates via REM showed an OR of
0.53 (95% CI, CL = 0.42–0.66) which was statistically
significant (p value < 0.01)
2. Studies defining early tracheostomy as that done within
14 days Pooling of estimates via REM showed an OR
of 0.78 (95% CI, CL = 0.67–0.9) which was statistically
significant (p value < 0.01).
3. Studies defining early tracheostomy as that done within
21 days Only one study showed an OR of 0.29 (95%
CI, CL = 0.12–0.75) with statistically significant p
value = 0.01.
regarding incidence of hospital acquired pneumonia.
Comparing the OR of the first two groups showed bet-
ter OR in studies defining early tracheostomy as that done
within 7 days (Fig. 1).
Under the random effects model, the point estimate and
95% confidence interval for the combined studies are 0.68
 Table 1  Data collected from the articles involved in comparing early and late tracheostomy in adults
Author Type of study Patients with early tracheostomy Patients with late tracheostomy
NO_ at (Days) HAP Mortality MV ICU stay NO_ at (Days) HAP Mortality MV ICU stay
Mean ± SD Mean ± SD Mean ± SD Mean ± SD

Ahmed and kuo [6] Retrospective cohort 29 < 7 11 4 15.7 ± 6 19 ± 7.7* 107 > 7 14 1 20 ± 16 25.8 ± 11.8
Alali et al. [7] Retrospective cohort 62 < 8 – 48 10 ± 5.65* 13 ± 5.65* 59 > 8 – 39 16 ± 6.36 19 ± 7
Alhajhusain et al. [8] Retrospective cohort 39 < 9 5* 5 15.1 ± 8.2* 16.6 ± 7.6* 63 > 9 25 23 27.2 ± 10.9 27.2 ± 9.1
Arabi et al. [9] Prospective cohort 29 < 7 – 5 9.6 ± 1.2* 10.9 ± 1.2* 107 > 7 – 15 18.7 ± 1.3 21 ± 1.3
Armstrong et al. [10] Retrospective cohort 62 < 7 – 7 – 15 ± 12* 95 > 7 – 11 – 29 ± 26
Barquist et al. [11] Prospective cohort 29 < 8 28 2 8.57 ± 7.9 4.96 ± 6 31 > 28 28 5 8.83 ± 9 5.26 ± 6.5
Ben-Avi et al. [12] Retrospective cohort 90 < 14 – 16* 16 ± 11 27 ± 21 109 > 14 – 17 21 ± 14 31 ± 23
European Archives of Oto-Rhino-Laryngology

Bickenbach et al. [13] Retrospective cohort 237 < 10 – 53* 20.1 ± 11.4* 25.5 ± 15.7* 59 > 10 – 24 29.5 ± 13.8 34.2 ± 24.1
Blot et al. [14] Prospective cohort 61 < 4 – 21 – – 95 > 14 – 20 – –
Bosel et al. [15] Prospective cohort 30 < 3 – 4 15 ± 4.94 17 ± 6.36 30 7–14 – 2 12 ± 5.65 18 ± 7
Chen et al. [16] Retrospective cohort 22 < 21 – – 44.4 ± 10.4* – 38 > 21 – – 59.5 ± 9.6 –
Choi et al. [17] Retrospective cohort 10 < 10 4* – 5.2 ± 6.5* 20.8 ± 6* 11 > 10 9 – 29.2 ± 22.9 38 ± 18.5
Devarajan et al. [18] Retrospective cohort 114 < 10 63* 24* – 526.5 ± 351.2* 114 14–28 71 46 – 698.7 ± 332.5
Flaatten et al. [19] Retrospective cohort 230 < 6 – 51* – – 231 > 6 – 75 – –
Ganuza et al. [20] Retrospective cohort 101 < 7 75 1 33.79 ± 29.62* 40.31 ± 26.49* 114 > 7 83 4 40.73 ± 22.71 49.42 ± 19.24
Gatti et al. [21] Prospective cohort 14 < 7 – – 12.6 ± 6.5* 15 ± 6.5* 10 > 7 – – 20.9 ± 10.7 27.1 ± 11.1
Gessler et al. [22] Retrospective cohort 39 < 7 19* 3* 17.4 ± 1.98* – 109 > 7 75 8 22.3 ± 1.77 –
Holloway et al. [23] Retrospective cohort 24 < 14 – 1 – 15 ± 7.5* 49 > 14 – 1 – 19 ± 11.8
Hossseinian et al. [24] Retrospective cohort 13,386 < 10 – 2837* – – 26,625 > 10 – 7147 – –
Huang et al. [25] Retrospective cohort 11 < 10 – 1* – 16 ± 8* 27 > 10 – 4 – 29 ± 13
Hyde et al. [26] Retrospective cohort 53 < 5 – – 16.7 ± 11.4* 21.4 ± 11* 53 > 5 – – 21.9 ± 12.9 28.6 ± 16.3
Jeon et al. [27] Retrospective cohort 39 < 10 2* 2 11.4 ± 5.6* 19.9 ± 10.6* 86 > 10 16 6 21.5 ± 15.5 31.1 ± 18.2
Keenan et al. [28] Retrospective cohort 5402 < 10 2479 568 – – 4260 > 10 2094 372* – –
Koch et al. [29] Retrospective cohort 50 < 4 19* 10 15.3 ± 5.4* 21.5 ± 7.5* 50 > 4 32 11 21.1 ± 7.24 30.6 ± 7.5
Mahafza et al. [30] Retrospective cohort 70 < 21 20* 13* – – 26 > 21 15 13 – –
Mehta et al. [31] Retrospective cohort 7595 < 7 – 831 – – 12,230 > 7 – 1850 – –
Moller et al. [32] Retrospective cohort 81 < 7 22* – 12.2 ± 0.9* 16.7 ± 1* 104 > 7 44 – 21.9 ± 1.3 26 ± 1.3
Pinheiro et al. [33] Retrospective cohort 11 < 8 6* 1* – – 17 > 8 12 8 – –
Puentes et al. [34] Retrospective cohort 32 < 7 – 8* – – 115 > 7 – 34 – –
Rizk et al. [35] Prospective cohort 1577 < 7 861* – – – 1527 > 7 1118 – – –
Rodrigues et al. [36] Prospective cohort 51 < 7 33* – 12 ± 1* 16 ± 1* 55 > 7 53 – 32 ± 3 37 ± 4
Rumback et al. [37] RCT​ 60 < 2 5* 19* 7.6 ± 4* 4.8 ± 1.4* 60 > 14 15 37 17.4 ± 5.3 16.2 ± 3.8
Scales et al. [38] Retrospective cohort 1081 < 9 – 426 – – 4773 > 13 – 1768 – –
Terragni et al. [39] RCT​ 145 6–8 30 – – – 119 13–15 44 – – –

13
 European Archives of Oto-Rhino-Laryngology

(0.60, 0.77). Using Trim and Fill, these values are unchanged

NO_ number of patients included in the study, at timing of tracheostomy, HAP number of patients developed hospital acquired pneumonia, MV duration of mechanical ventilation in days, ICU
26.27 ± 0.73
Mean ± SD
showing no publication bias.

20.8 ± 9.2
13.1 ± 8.2
33 ± 11.6
36.8 ± 26
ICU stay
Results of comparing early and late tracheostomy in adults

–
–
39.33 ± 1.33 regarding incidence of mortality:
Mean ± SD

16.7 ± 8.3
22.1 ± 7.6
32 studies were involved including 219,727 patients which
were divided into three groups according to the timing of
at (Days) HAP Mortality MV

early tracheostomy.
–
–
–
–

I-squared (I2) index was 83.36% denoting considerable

15,466

heterogeneity between studies.

Patients with late tracheostomy

141
894

Pooling of estimates via random effects method (REM)

56
71

–
4

showed that all studies regardless of the definition of early

3576
644

31
30
38

and late tracheostomy showed an odds ratio (OR) of 0.8

15

(95% CI, CL = 0.7–0.9) which was statistically significant (p

value < 0.01) favoring early tracheostomy over late tracheos-
> 10
> 10
> 10
> 10

> 3
> 7

tomy regarding incidence of mortality in adults.

15
71,244

259
61
312
50
7574
464
NO_

1. Studies defining early tracheostomy as that done within

7 days Pooling of estimates via REM showed an OR of
0.76 (95% CI, CL = 0.62–0.93) which was statistically
17.52 ± 1.38*

16.9 ± 13.2*
29.1 ± 24.4*
Mean ± SD

significant (p value 0.01).

13 ± 5.46
24 ± 9.6*
ICU stay

2. Studies defining early tracheostomy as that done within

14 days Pooling of estimates via REM showed an OR of
–
–

0.85 (95% CI, CL = 0.72–0.99) which was statistically

21.47 ± 1.86*

significant (p value 0.04).

Mean ± SD

13.3 ± 9.6*
14.9 ± 8.9*

3. Studies defining early tracheostomy as that done within

21 days Only one study showed an OR of 0.23 (95%
Mortality MV

CI, CL = 0.09–0.61) with statistically significant p

–
–
–
–

value < 0.01.
8947*
Patients with early tracheostomy

139
699

All favoring early tracheostomy over late tracheostomy

46
2*
8*

regarding incidence of mortality in adults.

2429*
347*
at (Days) HAP

Comparing the OR of the first two groups showed bet-

17*
29
7*
3*

ter OR in studies defining early tracheostomy as that done

within 7 days (Fig. 2).
16 < 10
Retrospective cohort 53,749 < 10
Retrospective cohort 5591 < 10

Under the random effects model, the point estimate and

256 < 3
315 < 4
128 < 7

58 3

95% confidence interval for the combined studies are 0.80

*Statistically significant difference between both groups

(0.70, 0.90). Using Trim and Fill, these values are unchanged
NO_

showing no publication bias.

Retrospective cohort
Retrospective cohort

Results of comparing early and late tracheostomy in adults

Type of study

regarding duration of mechanical ventilation:

Case series

stay duration of ICU stay in days

22 studies were involved including 4424 patients which were

RCT​
RCT​

divided into three groups according to the timing of early

tracheostomy. I-squared (I2) index was 98.79% denoting con-
Villwock and Jones [42]

siderable heterogeneity between studies.

Table 1  (continued)

Villwock et al. [41]

Pooling of estimates via random effects method (REM)

Young et al. [44]
Wang et al. [43]

Zagli et al. [46]

Tong et al. [40]

showed that all studies regardless of the definition of early

Yue et al. [45]

and late tracheostomy showed a standardized mean differ-

Author

ence (SMD) of − 1.06 (95% CI, CL = − 1.3 to − 0.82)

which was statistically significant (p value < 0.01) favoring

13
European Archives of Oto-Rhino-Laryngology

Table 2  Data collected from the articles involved in comparing early and late tracheostomy in pediatric age group
Author Type of study Patients with early tracheostomy Patients with late tracheostomy
NO_ at (Days) HAP MV ICU stay NO_ at (Days) HAP MV ICU stay
Mean ± SD Mean ± SD Mean ± SD Mean ± SD

Holscher et al. [47] Prospective cohort 11 < 7 2* 9 ± 4.5* 13 ± 6* 18 > 7 9 23 ± 5.5 25 ± 4.5
Lee et al. [48] Retrospective cohort 61 < 14 1 8.9 ± 3.66* 17 ± 13* 50 > 14 2 32.2 ± 21.25 42.5 ± 30.8

NO_ number of patients included in the study, at timing of tracheostomy, HAP number of patients developed hospital acquired pneumonia, MV
duration of mechanical ventilation in days, ICU stay duration of ICU stay in days
*Statistically significant difference between both groups

Fig. 1  Forest plot for the incidence of hospital-acquired pneumonia (HAP) in adults

early tracheostomy over late tracheostomy regarding dura- of − 2.94 (95% CI, CL = − 3.95 to − 1.93) which was
tion of mechanical ventilation in adults. statistically significant (p value < 0.01).
2. Studies defining early tracheostomy as that done within
1. Studies defining early tracheostomy as that done within 14 daysPooling of estimates via REM showed an SMD
7 daysPooling of estimates via REM showed an SMD

13

Fig. 2  Forest plot for mortality in adults
of − 0.82 (95% CI, CL = − 1.09 to − 0.55) which was
statistically significant (p value < 0.01).
3. Studies defining early tracheostomy as that done within
21 daysOnly one study showed an SMD of − 1.53 (95%
CI, CL = − 2.12 to − 0.93) with statistically significant
p value < 0.01.
All favoring early tracheostomy over late tracheostomy
regarding duration of mechanical ventilation in adults.
Comparing the SMD of the first two groups showed bet-
ter SMD in studies defining early tracheostomy as that done
within 7 days (Fig. 3).
13
European Archives of Oto-Rhino-Laryngology
Under the random effects model, the point estimate
and 95% confidence interval for the combined studies are
− 1.06 (− 1.3, − 0.82). Using Trim and Fill, these values are
unchanged showing no publication bias.
Results of comparing early and late tracheostomy in adults
regarding length of stay at the intensive care unit:
27 studies were involved including 143,506 patients which
were divided into two groups according to the timing of
early tracheostomy.I-squared (I2) index was 98.99% denoting
considerable heterogeneity between studies.
European Archives of Oto-Rhino-Laryngology
Fig. 3  Forest plot for the duration of mechanical ventilation (MV) in adults
Pooling of estimates via random effects method (REM)
showed that all studies regardless of the definition of early
and late tracheostomy showed a standardized mean differ-
ence (SMD) of − 0.82 (95% CI, CL = − 1.06 to − 0.58)
which was statistically significant (p value < 0.01) favor-
ing early tracheostomy over late tracheostomy regarding
length of stay at the ICU in adults.
1. Studies defining early tracheostomy as that done within
7 daysPooling of estimates via REM showed an SMD
of − 2.61 (95% CI, CL = − 3.44 to − 1.77) which was
statistically significant (p value < 0.01).
2. Studies defining early tracheostomy as that done within
14 daysPooling of estimates via REM showed an SMD
of − 0.66 (95% CI, CL = − 0.91 to − 0.41) which was
statistically significant (p value < 0.01).
Both favoring early tracheostomy over late tracheos-
tomy regarding length of stay at the ICU in adults.
Comparing the SMD of the two groups showed better
SMD in studies defining early tracheostomy as that done
within 7 days (Fig. 4).
Under the random effects model, the point estimate
and 95% confidence interval for the combined studies are
− 0.82 (− 1.06, − 0.58). Using Trim and Fill, these values
are unchanged showing no publication bias.
Meta‑analysis of studies involving pediatric patients
2 studies were involved including 140 patients. 11 patients
had early tracheostomy during the first week while 18
patients had late tracheostomy after the first week and 61
patients had early tracheostomy during the second week
while 50 patients had late tracheostomy after the second
week.
13

Fig. 4  Forest plot for the length of stay at the intensive care unit (ICU) in adults
Results of comparing early and late tracheostomy
in pediatric patients regarding hospital‑acquired
pneumonia:
I-squared (I 2) index was 0% denoting no heterogene-
ity between studies.Pooling of estimates via fixed effects
method (FEM) showed that all studies regardless of the
definition of early and late tracheostomy showed an odds
ratio (OR) of 0.27 (95% CI, CL = 0.06–1.15) which was
statistically insignificant (p value = 0.08).Therefore, early
tracheostomy is comparable to late tracheostomy regarding
incidence of hospital-acquired pneumonia in pediatric age
group (Fig. 5).
Results of comparing early and late tracheostomy
in pediatric patients regarding duration of mechanical
ventilation:
I-squared (I 2) index was 0% denoting no heterogeneity
between studies.Pooling of estimates via fixed effects
method (FEM) showed that all studies regardless of the
13
European Archives of Oto-Rhino-Laryngology
definition of early and late tracheostomy showed a stand-
ardized mean difference (SMD) of − 0.99 (95% CI, CL
= − 1.45 to − 0.52) which was statistically significant (p
value < 0.01) favoring early tracheostomy over late trache-
ostomy regarding duration of mechanical ventilation in
pediatric age group.
Results of comparing early and late tracheostomy
in pediatric patients regarding length of stay
at the intensive care unit:
I-squared (I 2) index was 81.27% denoting considerable
heterogeneity between studies. Pooling of estimates via
random effects method (REM) showed that all studies
regardless of the definition of early and late tracheostomy
showed a standardized mean difference (SMD) of − 1.65
(95% CI, CL = − 2.85 to − 0.46) which was statistically
significant (p value 0.01) favoring early tracheostomy over
late tracheostomy regarding length of stay at the ICU in
pediatric age group.
European Archives of Oto-Rhino-Laryngology
Fig. 5  Forest plot for the
incidence of hospital-acquired
pneumonia (HAP) in pediatric
age group
Discussion
Although tracheostomy is one of the most frequently per-
formed procedures in ICU patients due to its advantages
over prolonged endotracheal intubation, its timing is still a
controversy among different authors. There is limited evi-
dence to guide tracheostomy practice.
Many authors found that early tracheostomy reduces the
incidence of hospital acquired pneumonia (HAP) in adults
because of the reduced airway resistance, which acceler-
ates the weaning process of patients from the ventilator
decreasing tracheobronchial colonization and incidence of
ventilator-acquired pneumonia [6–22]. This agrees with our
results where the incidence of HAP was found to be signifi-
cantly less in patients who had an early tracheostomy done
when compared to those with late tracheostomy. Whereas
other authors stated that there is no difference between early
and late tracheostomy in the incidence of HAP [23–28].
This may be explained by many factors in different studies.
In the study done by Barquist et al. [24], there was a very
high incidence of pneumonia in both groups as most of the
patients included in the study acquired pneumonia (96.5%
in early and 90.3% in late tracheostomy groups) which may
have masked the differences between both groups. While
the study done by Ahmed and Kuo [23] had insufficient
number of patients that decreased the ability to observe a
significant difference between the two groups, Zagliet al.
[28] considered tracheostomy done after three days to be late
tracheostomy; this timing is still considered early by most
of the studies in this analysis. While in the study done by
Keenan et al. [26], all the participants were trauma patients
with different severity and all of them had respiratory failure
with unknown comorbidities.
Early tracheostomy reduces the rates of aspiration pneu-
monia, bacteremia, septic shock, ARDS, and multisystem
organ failure. As early tracheostomy decreases length of
stay in the ICU, it will decrease the incidence of these ICU-
associated complications (which may lead to multisystem
organ failure) leading to a reduction in the mortality rate if
compared to late tracheostomy in adults [8, 12, 14, 17, 20,
21, 29–35]. This agrees with our study where themortality
rate was significantly less in patients who had early trache-
ostomy compared to those who had late tracheostomy.
Other authors stated that there is no difference between
early and late tracheostomy regarding the mortality rate in
adults [6, 10, 11, 14, 18, 19, 23–25, 28, 36–43]. Again, the
study done by Ahmed and Kuo [23] had insufficient number
of patients that lead to the insignificant difference between
the two groups and Barquist et al. [24] had small sample
size. The low mortality rates reported in the studies done
by Arabi et al. [37], Blotet al. [39] and Guanzaet al. [25]
considerably reduced the probability of demonstrating the
effect of early tracheostomy on prognosis. Only one study
done by Keenanet al. [26] suggested that early tracheos-
tomy increases mortality rate. This study had a selection
bias regarding different types of injuries and unknown other
comorbidities. Study design, severity, and pattern of the dis-
order, as well as differences in the specialty of the ICUs,
caused the debate on the effect of early tracheostomy on
mortality.
Early tracheostomy decreases the need for sedation,
decreases airway resistance and allows more effective air-
way secretion suctioning and mouth care; it enhances patient
mobility, comfort and ability to speak and eat orally and pro-
vides more secure airway control through quicker and safer
tube replacement and fixation. This reduces the duration of
mechanical ventilation than those with late tracheostomy
in adults [6, 7, 9–11, 13, 16–18, 21, 25, 28, 30, 32, 36, 37,
44, 45]. This goes with our meta-analysis where the dura-
tion of mechanical ventilation was found to be significantly
13

less in patients who had an early tracheostomy done when
compared to those who had a late tracheostomy.
In spite of this, some authors found no effect of the timing
of tracheostomy on the duration of mechanical ventilation
[23, 24, 29, 40]. Ahmed and kuo [23] declared less ventila-
tion days with the early group (15.7 ± 6.0 versus 20.0 ± 16.0)
but statistically insignificant (p value = 0.570) which may be
due to the small number of patients contributed in the study.
The study done by Barquistet al. [24] had small sample size
and high rate of pneumonia (96.5% in early and 90.3% in late
tracheostomy groups) which may influence weaning from
mechanical ventilation. The study done by Ben-aviet al. [29]
was performed at a tertiary single centre; therefore, these
conclusions may not represent global practice. The study
done by Boselet al. [40] had heterogeneous small sample
size and generalizability was restricted by the single-centre
design.
Since early tracheostomy accelerates weaning from
mechanical ventilation and reduces aspiration and ventila-
tor-acquired pneumonias, then it enhances the possibility to
mobilize patients from the ICU and decreases length of ICU
stay than those with late tracheostomy in adults [6–8, 10,
11, 13, 16–21, 23, 25, 28, 30, 32, 34, 36–38, 45, 46]. In our
meta-analysis, the length of ICU stay was also found to be
significantly less in patients who had an early tracheostomy
done when compared to those with late tracheostomy.
While Younget al. [43], Barquistet al. [24], Boselet al.
[40] and Ben-aviet al. [29] found that the timing of trache-
ostomy did not affect the length of ICU stay, Younget al.
[43] only included patients at very high risk of prolonged
mechanical ventilation due to the absence of a predictive
tool to estimate the individual duration of mechanical venti-
lation, which affected recruitment to the study and the results
regarding ICU length of stay. In addition, the study did not
recruit its intended sample size due to recruitment fatigue
and exhaustion of funding. Again, the small sample size of
Boselet al. [40] and Barquistet al. [24] affected their results.
In pediatric age group, Holscheret al. [47] and Leeet al.
[48] both suggested a reduction in duration of mechanical
ventilation and length of ICU stay in patients with early tra-
cheostomy than those who had late tracheostomy due to the
reduction in the amount of sedatives required, increasing
patients’ comfort and allowing faster weaning from mechan-
ical ventilator by maintaining more favorable respiratory
conditions. However, both studies agreed about no signifi-
cant difference between both groups regarding incidence of
HAP as the incidence of post-tracheostomy pneumonia was
similar in both groups. There was a difference in defining
early tracheostomy between the two studies as Holscheret
al. [47] considered tracheostomy to be early if performed
before 7 days of intubation, while Leeet al. [48] considered
it early if performed before 14 days of intubation. Less num-
ber of studies found concerning tracheostomy in pediatric
13
European Archives of Oto-Rhino-Laryngology
age might be related to the infrequency of tracheostomy in
critically ill children as clinicians prefer more conservative
management of the pediatric airway due to the tracheos-
tomy-related complication rate which is approximately 39%
[49, 50].
This meta-analysis has some limitations. One of them is
the heterogeneity of the data which resulted from the dif-
ferent inclusion and exclusion criteria of each trial, varying
patients’ characteristics, tracheostomy techniques, and par-
ticularly the different definitions of early and late tracheos-
tomy.We have tried to overcome this heterogeneity through
doing a sensitivity analysis according to the methodology
of defining early tracheostomy but there were an inevitable
overlap between studies defining early tracheostomy within
7, 14, and 21 days of intubation. A second limitation of our
analysis is that not all randomized and quasi-randomized
trials that compare the outcomes of early tracheostomy with
late tracheostomy are included though a thorough search
was conducted to minimized missing relevant studies but
our search was limited to articles in English language, so,
articles in other languages may have been missed. A third
limitation is the inability to accurately predict the need for
prolonged mechanical ventilation on scientific bases so most
of the studies are based on the clinical judgement of the
investigators which may carry a selection bias. Another limi-
tation in assessing the pediatric age group was the inability
to analyze the mortality rates as there were no sufficient data
for conducting a meta-analysis due to paucity of studies.
Finally, the patients’ comorbidities and the reason for ICU
admission were not available from the studies and could not
be analyzed. Future studies on the effect of different trache-
ostomy techniques on the morbidity in early and late tra-
cheostomies are needed to guide us for the better technique.
Conclusion
Our meta-analysis suggests that in adult patients with pro-
longed intubation, early tracheostomy is significantly asso-
ciated with reduction in incidence of hospital acquired
pneumonia, incidence of mortality, duration on mechani-
cal ventilation and length of stay in ICU, especially when
early tracheostomy is performed within the first 7 days of
intubation (according to our sensitivity analysis), while in
pediatric age group, early tracheostomy is associated with
decreased duration of mechanical ventilation and length of
stay in ICU but does not affect the incidence of hospital
acquired pneumonia.
Compliance with ethical standards 
Funding None
European Archives of Oto-Rhino-Laryngology
Conflict of interest  The authors declare that they have no conflict of
interest. The entire work was approved from Ain Shmas University
ethical committee.
Ethical approval  This article does not contain any studies with human
participants or animals performed by any of the authors.
References
1. Frost E (1976) Tracing the tracheostomy. Ann Otolaryngol
85:618–624
2. Esteban A, Anzueto A, Alia I et al (2000) How is mechanical
ventilation employed in the intensive care unit? An international
utilization review. Am J Respir Crit Care Med 161:1450–1458
3. Plummer A, Gracey D (1989) Consensus conference on artifi-
cial airways in patients receiving mechanical ventilation. Chest
96(1):178–1780
4. Kane T, Rodriguez J, Luchette F (1997) Early versus late trache-
ostomy in the trauma patient. Respir Care Clin 3:1–20
5. Maziak D, Meade M, Todd T (1998) The timing of tracheotomy
a systematic review. Chest 114:605–609
6. Alhajhusain A, Ali A, Najmuddin A et  al (2014) Timing of
tracheotomy in mechanically ventilated critically ill mor-
bidly obese patients. Crit Care Res Pract 2014. https://doi.
org/10.1155/2014/840638
7. Choi HJ, Paeng SH, Kim ST et al (2013) The effectiveness of
early tracheostomy (within at least 10 days) in cervical spinal cord
injury patients. J Korean Neurosurg Soc 54(3):220–224
8. Devarajan J, Vydyanathan A, Xu M et al (2012) Early trache-
ostomy is associated with improved outcomes in patients who
require prolonged mechanical ventilation after cardiac surgery. J
Am Coll Surg 214:1008–1016
9. Gessler F, Mutlak H, Lamb S et al (2015) The impact of tracheos-
tomy timing on clinical outcome and adverse events in poor-grade
subarachnoid hemorrhage. Crit Care Med 43:2429–2438
10. Jeon Y, Hwang J, Lim Y et al (2014) Effect of tracheostomy tim-
ing on clinical outcome in neurosurgical patients: early versus late
tracheostomy. J Neurosurg Anesthesiol 26:22–26
11. Koch T, Hecker B, Hecker A et al (2012) Early tracheostomy
decreases ventilation time but has no impact on mortality of inten-
sive care patients: a randomized study. Langenbeck’s Arch Surg
397(6):1001–1008
12. Mahafza T, Batarseh S, Bsoul N et al (2012) Early vs. late tra-
cheostomy for the ICU patients: experience in a referral hospital.
Saudi J Anaesth 6:152–154
13. Moller M, Slaikeu J, Bonelli P et al (2005) Early tracheostomy
versus late tracheostomy in the surgical intensive care unit. Am J
Surg 189(3):293–296
14. Pinheiro V, Tostes R, Brum C et al (2010) Early vs late tracheos-
tomy in patients with acute severe brain injury. J Bras Pneumol
36:84–91
15. Rizk E, Patel A, Stetter C et al (2011) Impact of tracheostomy
timing on outcome after severe head injury. Neurocrit Care
15:481–489
16. Rodriguez J, Steinberg S, Luchetti F et al (1990) Early tracheos-
tomy for primary airway management in the surgical critical care
setting. Surgery 108(4):655–659
17. Rumback M, Newton M, Truncale T et al (2004) A prospec-
tive, randomized study comparing early percutaneous dilational
tracheotomy to prolonged translaryngeal intubation (delayed
tracheotomy) in critically ill medical patients. Crit Care Med
32(8):1689–1694
18. Tong C, Kleinberger A, Paolino J et al (2012) Tracheotomy tim-
ing and outcomes in the critically Ill. Otolaryngol–Head Neck
Surg 147(1): 44–51
19. Villwock J, Villwock M, Deshaies E (2014) Tracheostomy
timing affects stroke recovery. J Stroke Cerebrovasc Dis
23:1069–1072
20. Villwock J, Jones K (2014) Outcomes of early versus late trache-
ostomy: 2008–2010. Laryngoscope 124:1801–1806
21. Wang H, Lu K, Liliang P et al (2012) The impact of tracheostomy
timing in patients with severe head injury: an observational cohort
study. Injury 43:1432–1436
22. Yue Z, Sui F, Chen XK et al (2012) Early versus late percutane-
ous dilational tracheostomy in critically ill patients anticipated
requiring prolonged mechanical ventilation. Chin Med J (Engl)
125:1925–1930
23. Ahmed N, Kuo Y (2007) Early versus late tracheostomy in patients
with severe traumatic head injury. Surg Infect 8(3):343–347
24. Barquist E, Amortegui J, Hallal A et al (2006) Tracheostomy in
ventilator dependent trauma patients: a prospective, randomized
intention to treat study. J Trauma 60(1):91–97
25. Ganuza J, Garcia Forcada A, Gambarrutta C et al (2011) Effect
of technique and timing of tracheostomy in patients with acute
traumatic spinal cord injury undergoing mechanical ventilation.
J Spinal Cord Med 34(1):76–84
26. Keenan J, Gulack B, Nussbaum D et al (2015) Optimal timing of
tracheostomy after trauma without associated head injury. J Surg
Res 198(2): 475–481
27. Terragni P, Antonelli M, Fumagalli R et al (2010) Early vs late
tracheotomy for prevention of pneumonia in mechanically ven-
tilated adult ICU patients: a randomized controlled trial. JAMA
303(15):1483–1489
28. Zagli G, Linden M, Spina R et al (2010) Early tracheostomy
in intensive care unit: a retrospective study of 506 cases of
video-guided Ciaglia Blue Rhino tracheostomies. J Trauma
68(2):367–372
29. Ben-Avi R, Ben-Nun A, Levin S et al (2014) Tracheostomy after
cardiac surgery: timing of tracheostomy as a risk factor for mortal-
ity. J Cardiothorac Vasc Anesth 28:493–496
30. Bickenbach J, Fries M, Offermanns V et al (2011) Impact of early
vs. late tracheostomy on weaning: a retrospective analysis. Min-
erva Anestesiol 77:1176–1183
31. Flaatten H, Gjerde S, Heimdal JH et al (2005) The effect of tra-
cheostomy on outcome in intensive care unit patients. Acta Anaes-
thesiol Scand 50:92–98
32. Gatti C, Cardu C, Bentini C et al (2004) Weaning from ventilator
after cardiac operation using the Ciaglia percutaneous tracheos-
tomy. Eur J Cardiothorac Surg 25:541–547
33. Hosseinian L, Chianget Y, Itagakial S et al (2013) Earlier versus
later tracheostomy in patients with respiratory failure after car-
diac surgery in the United States. J Cardiothorac Vasc Anaesth
28(3):488–492
34. Huang Y, Lee T, Liao C et al (2013) Tracheostomy in craniec-
tomised survivors after traumatic brain injury: a cross-sectional
analytical study. Injury 44:1226–1231
35. Puentes W, Jerath A, Djaiani G et al (2016) Early versus late tra-
cheostomy in cardiovascular intensive care patients. Anaesthesiol
Intensive Ther 48(2):89–94
36. Alali A, Scales D, Fowler R et al (2014) Tracheostomy timing
in traumatic brain injury: a propensity-matched cohort study. J
Trauma Acute Care Surg 76:70–76
37. Arabi Y, Haddad S, Shirawi N et al (2004) Early tracheostomy
in intensive care trauma patients improves resource utilization: a
cohort study and literature review. Critical Care 8(5):347–352
38. Armstrong P, McCarthy M, Peoples J (1998) Reduced use of
resources by early tracheostomy in ventilator-dependent patients
with blunt trauma. Surgery 124(4):763–766
13

39. Blot F, Similowski T, Trouillet J et al (2008) Early tracheotomy
versus prolonged endotracheal intubation in unselected severely
ill ICU patients. Intensive Care Med 34(10):1779–1787
40. Bösel J, Schiller P, Hook Y et al (2012) Stroke-related early
tracheostomy versus prolonged orotracheal intubation in neuro-
critical care trial (SETPOINT): a randomized pilot trial. Stroke
44(1):21–28
41. Mehta A, Cooke C, Wiener R et al (2016) Hospital variation in
early tracheostomy in the United States: a population-based study.
Crit Care Med 44(8):1506–1514
42. Scales D, Thiruchelvam D, Kiss A et al (2008) The effect of tra-
cheostomy timing during critical illness on long-term survival.
Critic Care Med 36(9):2547–2557
43. Young D, Harrison D, Cuthbertson B et al (2013) Effect of early
vs late tracheostomy placement on survival in patients receiving
mechanical ventilation: the TracMan randomized trial. Caring for
the critically Ill patient. JAMA 309(20):2121–2129
44. Chen H, Song L., Chang H et al (2016) Factors related to trache-
ostomy timing and ventilator weaning: findings from a population
in Northern Taiwan. Clin Respir J 1–8
13
European Archives of Oto-Rhino-Laryngology
45. Hyde G, Savage S, Zarzaur B et al (2014) Early tracheostomy in
trauma patients saves time and money. Injury 46(1):110–114
46. Holloway A, Spaeder M, Basu S (2015) Association of timing of
tracheostomy on clinical outcomes in PICU patients. Paediatric
Criti Care Med 16(3):52–58
47. Holscher C, Stewart C, Peltz E et al (2014) Early tracheostomy
improves outcomes in severely injured children and adolescents.
J Pediatr Surg 49:590–592
48. Lee J, Koo C, Lee S et al (2016) Effect of early vs. late tracheos-
tomy on clinical outcomes in critically ill pediatric patients. Acta
Anaesthesiol Scand 1281–1288
49. Wood D, McShane P, Davis P (2012) Tracheostomy in children
admitted to paediatric intensive care. Arch Dis Child 97: 866–869
50. Pereira K, Macgregor A, Mcduffie C et al (2003) Tracheostomy
in preterm infants: current trends. Arch Otolaryngol Head Neck
Surg 129:1268–1271