TheInjury,

JournalInfection,
ofTRAUMAand Critical Care

Early Tracheostomy versus Prolonged

Intubation in Severe Head Injury
Endotracheal
Moulay Ahmed Bouderka, Bouchra Fakhir, Abderrahmane Bouaggad, Badreddine
Hmamouchi,
Driss Hamoudi, and Abdeslam Harti

Background: To see if early trache- longed endotracheal intubation

After pneumonia
group (I was diagnosed, mechan-
ostomy (fifth day) reduces duration of me- group, n 31). We evaluated total
ical ventilatory
time time was 6 4.7 days for
chanical ventilation, ICU stay, incidence of mechanical ventilation, ICU ETstay,
group versus 11.7 6.7 days for PEI
of pneumonia and mortality in compari- pneumonia incidence and mortality. 0.01). Theregroup was no( pdifference
son with prolonged intubation (PI) in pa- Complications related to eachin frequency
techniqueof pneumonia or mortality
tients with head injury. were noted. Analysis of data were per- between the two groups.
Methods: Patients were prospec- formed using Yates and Kruskall Walis Conclusion:
In severe head injury
tively included in this study if they met the tests. p < 0.05 was considered
early
significant.
tracheostomy decreases total days
following criteria: isolated head injury, Results: The two groups
of mechanical
were com-ventilation or mechanical
Glasgow coma scale (GCS) score < 8on parable in term of age, sex, andventilation
Simplifiedtime after development of
first and fifth day, with cerebral contusion Acute Physiologic Score (SAPS). The pneumonia.
on CT scan. On the fifth day, randomiza- mean time of mechanical ventilatory sup- Tracheostomy, Key words:
Head in-
tion was done in two groups: early trache- port was shorter in T group
jury,(14.5
Intubation,
7.3) Nosocomial pneumonia,
ostomy group (T group, n 31) and pro- versus I group (17.5 10.6) ( p 0.02). Intensive care unit.
2004;57:251254.
J Trauma.

T racheostomy is frequently done in ICU. ManyThis

ommend it to avoid serious oropharyngeal
jury occurring from prolonged translaryngeal
early
authorsstudy
andtracheostomy
rec-
was conducted to further define the impact of
larynx in- (T) on duration of mechanical ventilation,
ICUintubation.
stay, nosocomial
1 7 pneumonia and mortality in compari-
However, the benefit of tracheostomy in the ICU is son
notwithclearly
prolonged endotracheal intubation (I) in patients
defined. Despite the long history of tracheostomy few data are with isolated severe head injury.
available to define the impact of early tracheostomy on duration
of mechanical ventilation and ICU stay. In patients receiving MATERIALS AND METHODS
mechanical ventilation, tracheostomy has the followingThis benefits:
is a prospective randomized study conducted 2
easier nursing care became easier, improved comfort, yearsmoreafter local CRB committee approval. Patients were
secure tube with increased patient mobility, allowance included
of in the study if they met the following criteria:
speech, oral nutrition and in some studies early weaning Isolated
from severe head injury (admission Glasgow coma
mechanical ventilation.8, 9 Conversely, some studies have sug- scale (GCS) score 8).
gested that tracheostomy is associated with an increased risk of Cerebral contusion on CT scan
nosocomial pneumonia. 1 0 1 2 GCS score 8 on the fifth day without any sedation.
In 1989 the American consensus conference on On artificial
the fifth day of hospitalization if these criteria were
airways issued the statement that tracheostomypresent, is preferred
patientsif were randomized to: early tracheostomy (T)
the need for an artificial airway is anticipated to beorgreater prolonged endotracheal intubation (I). In the T Group,
than 21 days. 4 In 1998 the European consensus arrived tracheostomy
at the was performed on the 5th or 6th day after
same conclusion. 1 3 admission using a standard technique in the ICU by a critical
care physician with low pressure Tracheostomy tube cuffs.
We compared demographic date, admission scores and
outcome, specifically evaluation the time of mechanical ven-
Submitted for publication January 24, 2003.
Accepted for publication June 30, 2003. tilation, Simplified Acute Physiologic Score (SAPS), ICU
Copyright 2004 by Lippincott Williams & Wilkins, Inc. stay, frequency of pneumonia and Clinical
mortality.
but
14

Volume
Ibn From57 for
Anesthesiologist
thesiology
Morocco;
Rochd
Manuscript
Address
DOI:email:
and Number
Hospital,
the 2in Unit
10.1097/01.TA.0000087646.68382.9A
Intensive
Department
mabouderka@yahoo.fr.
(ASA)
published
reprints:
Casablanca,
Annual
Care
Moulay
of Anesthesiology
abstract
Meeting,
Morocco.
Ahmed
(P33),
formOctober
Ibn
Bouderka,
inRochd
and
The2001. not
Intensive
American 6.01.
Hospital, prolonged
Department
Careendoscopic
All ofOur
Society data
Casablanca,
UnitAnes-hypothesis
were
of(P33), Statistics
endotracheal
diagnosis
complications
recorded
was
of that
in
intubation
pneumonia
arelated
PC
early
and
tracheostomy
to
(PEI)
was
analyzed
tracheostomy
by
were
CDC
by
noted.
could
Epi
criteria.
andinfo
The
251
re-
to
15
 TheInjury,
JournalInfection,
ofTRAUMAand Critical Care

duce the mechanical ventilation time by 25%. We group.

fixedGram-negative
risk bacilli, especially Acinetobacter and
at 5% and risk at 20%. So the minimal number Pseudomonas,
for our were the most frequent bacteria isolated in
population study was 34. Randomization was done nosocomial
with haz-pneumonia (Fig. 1). There were two non-lethal
ard table permutation. stomal bleeds and five stomal infections in T group. Inspira-
The characteristics of patients in each groups
torywere
dyspnea
com-was present in one patient in T group and three
pared using Chi 2 test with Fischers exact testpatients
and Yatesin test
I group. Laryngotracheal endoscopic examination
correction when necessary for qualitative variables. found
We com-one tracheal stenosis in the T group that required
pared continuous variables using Students t test.
surgical
p 0.05
treatment and five inflammatory granuloma in I
was considered significant. group with good recovery after treatment with corticoste-
roids. There was no difference in mortality between the two
RESULTS groups (Table 4). Intracranial hypertension, acute respiratory
A total of 150 patients with severe head injury
distress
hadsyndrome
been (ARDS) and sepsis were the major causes
admitted to our ICU during the study period. Among these of death for the two groups (Table 5).
patients, six were excluded for the following reasons: 1)
improvement of GCS score 8 on the fifth day (n 2); 2) DISCUSSION
death during the first week of hospitalization (n 4). It is a paradox that although tracheostomy is frequently
The two groups were comparable in terms ofrecommended
age, sex, in head injury patients, there are few studies
and SAPS (Table 1). The mean time of mechanicalstrictlyventila-
related to this group. There has been a little agreement
tory support was shorter in T group than in PEI group (Table
2). Nosocomial pneumonia was not more frequent in ET
group than in I group (Table3). However the day pneumonia
was diagnosed was later for T group than for PEI group. After
nosocomial pneumonia has been acquired, the number of
mechanically ventilated days was greater in I group than in T

Table 1 Demographic Data

T Group I Group
p
(n 31) (n 31)

Age (years) 41.1 17.5 40 19 0.53

Sex: M/F 18/9 20/11
SAPS 5.4 1.5 6 3.8 0.52
F, female; M, male; SAPS, simplified acute physiological score.

Table 2 Ventilatory Data Bacteria isolatedFig.

during
1. nosocomial pneumonia (KES, kleb-
ssiela, enterobacter, serratia mercesens; PGC, positives gram cocci).
T Group I Group
p
(n 31) (n 31)

Number of weaning attempts 1.6 0.7 1.5 0.9 0.6 Table 4 Outcome and Mortality
Total ventilation days 14.5 7.3 17.5 10.6 0.02
Day of extubation or 26.3 13.7 19.4 10.4 0.03 T Group I Group
p
(n 31) (n 31)
tracheostomy off
Recovery n (%) 19 (61.3) 23 (74.2) 0.41
Table 3 Frequency of Nosocomial Pneumonia and Bleeding n (%) 2 (6.4) 0 (0) 0.47
Death n (%) 12 (38.7) 7 (22.5) 0.27
Sinusitis
Day of death 30.6 20.1 27.7 14.6 0.72
T Group I Group
p
(n 31) (n 31)

Nosocomial pneumonia n (%) 18 (58) 19 (61.3) 0.79 Table 5 Etiology of Death

252
positives
Day
Mechanically
Bacteria
NGB/PGC
Sinusitis
NP
pneumonia
NP, diagnosed
gram
nosocomial
n (%)
13/5
ventilated
bacilli.
diagnosed
315/4
6.7
(9.6)
pneumonia;
1.8
0.92
days
5 9.2
(16.1)
after
2.3 0.7
NGB,
0.95 negatives
6 4.7
gram
11.7
bacilli;
6.7 0.01
PGB, Intracranial
Sepsis
ARDS
Cardiovascular
ARDS, 17 32acute
hypertension
respiratory
dysfunction 10August
4distress TI2004
1 syndrome.
Group
(n 31)
 Tracheostomy versus Intubation in Severe Head Injury

on the optimal timing of the procedure in mechanically protection. ven-Early tracheostomy may assist in early termina-
tilated patients. 1 6 1 8 The consensus conferencetion of 1989
of mechanical
rec- ventilatory support and therefore, reduce
ommended conversion to tracheotomy if the anticipated need the ICU and hospital stay for these patients.
for mechanical ventilation is 21 days. 4 Few studies have
In our experience, two clinical features at the time of
tried to resolve this question: is there any benefitintubation to practice aid identification of unlikely to be extubated: pa-
tracheotomy? If yes, at what time? tients with a GCS score 8, brain stem deficit and cerebral
Few studies methodologically acceptable tried to answer
contusion in CT scan. That is why we chose these conditions
these questions. 8 , 9 , 19 2 2 Rodriguez et al. suggested
8 that therecriteria in our study. The limitation of this ap-
as inclusion
was a reduction in the duration of mechanical ventilation, proach is the high mortality rate during the first week of
ICU and hospital stay. He did not comment on sequellea of
hospitalization. In our opinion tracheostomy should depend
prolonged intubation versus early tracheostomy.not Dunham et
only on the probability of successful extubation but also
al.1 9 found no difference with respect to importantonclinical the probability that a patient had good chance of being
outcomes. However, they noted that the frequency discharged of major alive. That is why we avoided in our protocol
laryngotracheal damage was similar for early andtracheostomy late trache- in patients who will be potentially extubated
ostomy. El-Naggard et al. 2 0 found a higher percentage
early (GCS of score 8) and the patient who would die during
extubation and lower frequency of airway lesions inthethefirst lateweek and would not benefit from the procedure.
tracheotomy groups. Two retrospective studies were Presently, con- most physicians decide the timing of trache-
ducted by Lesnik et al. and Blot et al. Lesnikostomy
8 21 reviewed 101
in patients with neurologic disease based on results
adult patients who were admitted after blunt injuries, 32 had
derived from observations in patients mechanically ventilated
tracheostomy within the first 4 days and 69 underwent tra- for pulmonary
Most ofcauses. the patients
3 , 2 3 with sever
cheostomy after 4 days. The author found that the mean
head injury, as seen in our study, required intubation for at
duration of ventilatory support was 6.0 days in early trache-
least one week but not necessarily mechanically ventilation
ostomy group versus 20.6 days in the late tracheotomy group
for a long time, However, airway protection is permanently
(p 0.001). Blot et al. 2 1 in their retrospective study com-
needed. Tracheostomy provides an early alternative for air-
pared early (with in 48 h) versus late ( 7days) tracheostomy
way protection and seems to decrease the need for prolonged
in 53 neutropenic patients. The frequency of nosocomial
mechanical ventilation support. Secondly, severe head injury
pneumonia, death in ICU and hospital were not significantly
patient requires a long time for recovery and the airways
different but the length of the hospital stay and of the me-
protective reflexes are rarely optimal. This was confirmed by
chanical ventilation were significantly longer in early trache-
the high frequency of reintubation attributable to poor control
ostomy group ( p 0.05).
over secretions aspiration in intubation groups. 2 4 ,2 5
However all these studies have some methodological
Frequency of nosocomial pneumonia seems to be lower
bias:
with early tracheostomy. Many studies confirm theses
Inhomogeneous population 1 8 -2 0
Other
findings.
authors
9 , 2 found
1 ,2 6 that tracheostomy in-
Retrospective study 9 , 2 0
No randomization 9 , 1 8 creases the frequency Weofdidpneumonia.
not find2 any
7 2 9

If randomization was present, it was substandard, statistical difference in term of frequency of pneumonia be-
(alter-
native allocation, the day of the studies were tween the 2 groups.
blinded)8 , 1 8 ,1 9 The association between the risk of laryngotracheal in-
Comparison between early and late tracheotomy jury andnotduration
ver- of intubation is another important consid-
sus prolonged endotracheal intubation 8 , 9, 1 7 21 eration in the timing of tracheostomy. Nowak reported
et al.30

Thats why specific inclusion and exclusionthat the risk

mogenous population and standardized weaning patients
necessary. We think that to ovoid all these methodological
bias, we have to find first a homogenous population
of pathology and in terms of gravity. Second, thetween
criteria, ho- of severe tracheal complications was higher in
practiceswith are closed head injury who were intubated for 14
days compared with those intubated for 14 days and there
wasinno termdifference in the risk of laryngotracheal injury be-
06 days and 713 days of intubation. Richard et al.
population 31

chosen must be ventilated for a long time in orderevaluated not to dothe frequency rate of laryngotracheal stenosis in
tracheostomy abusively (for patients who did not315 patients
require it). with neurologic disease (Head injury, tetraple-
We believe that the ideal population study is patients gia. .with
.), the risk of laryngotracheal stenosis was not depen-
isolate
methodology.
who
unsuccessful
the
ical
Volume
first observed
ventilatory
We
were
severe
57
week,
had
mechanically
extubation
Number
head
most
support injury
of2that
these
but
ventilated
and
who
patients
were
patients
required
metintubated
had
with
the
didnt
tracheostomies.
criteria
asever
high
mainly
require
performed
frequency
complications.
difference
head
ical
cited
was
dent
cheal
symptoms
for
mechan-
injury
in
done
on
After
airway
complication.
our
before
of
the
between
only
duration
were
However,
if
thethere
the
recorded
time
We
of
two
was
we
intubation.
period
think
groups
should
clinical
and that
when
laryngotracheal
inremember
Tracheostomy
Our
suspicion
term
laryngotracheal
study
of laryngotracheal
that
of
didlaryngotra-
endoscopy
should
only
not injury
find
clin-
253
bea
 TheInjury,
JournalInfection,
ofTRAUMAand Critical Care

becomes a concern, based on experience from other studies 2 9 Le

14. 31 Gall JR, Loirat P, Alperovitch A, et al. Simplified acute

of patients with neurologic diseases. physiology score for ICU1984;12:975977.

patients.
Crit Care Med.
15. Garner
Kollef32 found that patient with respiratory failure who JS, Jarvis WR, Emori TG, Horan TC, Hughes JM. CDC
definitions for nosocomial Aminfections.
J infect control
1988;
had a tracheostomy had more favorable outcome than patients 16:128140.
who did not have a tracheostomy. However, this study pop- DE, Meade MO, Todd TRJ. The timing of tracheotomy. A
16. Maziak
ulation was not homogenous and they included all of their systemic review. Chest.1998;114:605609.
17. Lewis
ICU patients. In our study mortality did not differ between RJ. Tracheostomies: indications, timing, and complications.
the two groups. Clin Chest Med 1992;13:137149.
18. Gibbons KJ. Tracheostomy. Timing Crit
is everything.
Care Med.
In conclusion, early tracheotomy decreases the total days 2000;28:16631664.
of the ventilator and mechanical ventilatory time after19.
pneu-
Durham CM, Lamonica C. Prolonged tracheal intubation in the
monia in isolated severe head injury. It did not reduce either trauma patient. J Trauma. 1984;24:120124.
ICU stay or the frequency of pneumonia or mortality. 20. El Naggar M, Sadagopan S, Levine H, et al. Factors influencing
choice between tracheostomy and prolonged translaryngeal
intubation in acute respiratory failure: a prospectiveAnesth
study.
Analg. 1976;55:195201.
REFERENCES 21. Blot F, Guiguet M, Antonn, S, et al. Early tracheotomy in
1. Colice GL, Stukel TA, Dain B. Laryngeal complications of neutropenic, mechanically ventilated patients: rational and results of
prolonged intubation. Chest. 1989;96:877884. a pilot study. Support Care 1995;3:291296.
Cancer.
2. Bishop MJ. The timing of tracheotomy: an evolving consensus. 22. Sugerman HJ, Wolfe L, Pasquale MD, et al. Multicenter,
Chest. 1989;96:71213. randomized, prospective trial of early tracheostomy.
J Trauma.
1997;
3. Heffner JE. Timing of tracheotomy in mechanically ventilated 43:741747.
patients. Am Rev Respir Dis. 1993;147:768771. 23. Heffner JE, Miller KS, Salon SA. Tracheostomy in the intensive
4. Plummer AL, Gracey DR. Consensus conference on an artificial care unit. Part I: indications, technique management.
Chest.
1986;
airway in patients receiving mechanical ventilation. Chest. 1989; 90:269274.
96:178180. 24. Koh WY, Lew TW, Chin NM, et al. Tracheostomy in a neuro-
5. Stauffer JL, Olson DE, Petty TL. Complications and consequences intensive care setting: indications
Anesth Intensive
and timing.
Care.
of endotracheal intubation and tracheotomy: A prospective study of 1997;25:365368.
150 critically ill adult patients. Am J Med. 1981;70:6576.25. Qureshi AI, Suarez JI, Parekh PD, Bharwej A. Prediction and timing
6. Whited RE. A prospective study of laryngotracheal sequelae in long-of tracheostomy in patients with infratentorial lesions requiring
term intubation. Laryngoscope. 1984;94:367377. mechanical ventilatory2000;28:13831387.
support.
Crit Care Med.
7. Austin RT. Tracheostomy and prolonged intubation in the 26. Kluger Y, Paul DB, Lucke J, et al. Early tracheostomy in trauma
management of trauma. Injury. 1971;2:191198. patients. Euro J Emerg Med1996;3:95101.
8. Rodriguez JL, Steinberg SM, Luchetti FA, et al. Early tracheostomy 27. Georges H, Leroy O, Guery B, Alfandari S, Beaucaire G.
for primary airway management in surgical critical care setting.
Predisposing factors for nosocomial pneumonia in patients receiving
Surgery. 1990;108:655659. mechanical ventilation and requiring tracheotomy.
Chest.
2000;
9. Lesnik I, Rappaport W, Fulginiti J, et al. The role of early 118:767774.
tracheostomy in blunt, multiple organ trauma. Ann Surg. 1992; 28. Artigas AT, Dronda SB, Valles EC, et al. Risk factors for
58:346349. nosocomial pneumonia in critically ill traumaCrit patients.
Care
10. Celis R, Torres A, Gelett J, et al. Nosocomial pneumonia: a Med. 2001;29:304309.
multivariate analysis of risk and prognosis. Chest. 1988;93:318324.
29. Heffner JE, Scott MK, Sahn SA. Tracheostomy in the intensive care
11. Gross AS, Roub B. Role of respiratory assistance devices in endemic unit: complications. Chest. 1986;90:430435.
nosocomial pneumonia. Am J Med. 1981;70:681685. 30. Nowak P, Cohn AM, Guidice MA. Airway complications in patients
12. Kollef MH, Von Harz B, Prentice D, et al. Patient transport from with closed-head injuries.Am J1987;8:9196.
Otolaryngol.
intensive care increases the risk of developing ventilation associated
31. Richard I, Giraud M, Perrouin-Verbe B, et al. Laryngotracheal
pneumonia. Chest. 1997;112:765773. stenosis after intubation or tracheostomy in patients with
13. Chastre J, Bedock B, Clair B, et al. Quel abord trachal pour la neurological disease.Arch Phys
1996;77:493496.
Med Rehabil.
ventilation mcanique des malades de ranimation ? VIIIme 32. Kollef MH, Ahrens TS, Shannon W. Clinical predictors and
confrence consensus en ranimation et mdecine durgence. outcomes Ran for patients requiring tracheostomy in the intensive care
Urg. 1998;7:435442. unit. Crit Care Med. 1999;27:17141720.

254 August 2004