Content uploaded by Serdar Kahraman
Author content
All content in this area was uploaded by Serdar Kahraman on Feb 09, 2014
Content may be subject to copyright.
Neurosurg.
Rev.
20 (1997) 17%181
Causes of infections and management results in penetrating
craniocerebral injuries
Engin G/Jniil, Alper Baysefer, Serdar Kahraman, Ozcan ~lklatekerlio~lu, Ferruh
Gezen, Orhan Yayla, and Naci Seber
Department of Neurosurgery, Faculty of Medicine, GATA, Ankara, Turkey
Abstract
From February 1992 to December 1994, 148 patients with
penetrating craniocerebral injuries we~ze treated surgically
with primary and secondary debridement including repair
of dural defects and removal of retained intracranial bone
and metal fragments. Dural defects were closed primarily
or with temporalis fascia, pericranium, and cadaver graft.
Cerebrospinal fluid fistulas were observed in 11 (7.3 %) pa-
tients; 7 of these were infected. Central nervous system
(CNS) infection was seen in 2 patients without CSF fistula.
Excluding those 11 patients with CSF fistula, CNS infec-
tion was shown in 2 of the 137 cases (1.5 %). All patients
underwent CT scans periodically. In 51 (34 %) of 148 pa-
tients, bone aM metal fragments were determined on control
CT scans. During this time, 12 patients died (8 %). Most of
deaths were caused by the direct effect of brain injury and
occured within the first month after injury. Fragments re-
tained after first debr_idement were followed periodically
by CT scan. Surgery was not performed until infection de-
veloped. Retained fragments did not increase the infection
risk, but high rates of infection did occur in cases with CSF
fistula.
Keywords:
Craniocerebral injury, penetrating wound, re-
tained bone fragments, surgical infection.
sity of intensive antibiotic therapy after extensive re-
section as well as the removal of infectious sources
such as retained bone and metal fragments in the
follow-up period. At this time, a significant decrease
of the mortality rates was observed. In later experi-
ence as in some recent reports, mortality rates are
stated to be low because of use of new generation
antibiotics and less aggressive surgery. The mortal-
ity and morbidity in craniocerebral missile wounds
are affected by many factors, including transport of
the patient, antibiotic therapy, techniques of surgical
approaches, and follow-up procedures [1, 3, 5, 7, 15].
In this study we consider that adequate initial de-
bridement with removal of fragments not located
deeply, arrangement of measures to prevent con-
tamination via CSF fistula, and follow-up with peri-
odic CT scans is sufficient in the management of cra-
niocerebral missile wounds. Since retained frag-
ments do not increase the rate of infection, avoiding
secondary operations which have the risk of aggra-
vating the neurological deficits should be beneficial
[1, 3, 4, 8, 161.
1 Introduction
Neurosurgical experience with craniocerebral gun-
shot injuries has been reported since World War I.
Therefore, mortality rates of penetrating head
wound has decreased recently. In the early experi-
ence less aggressive surgical approaches and a long
transport time resulted in higher infection rates and
increased mortality [2]. At the time of World War I,
neurosurgeons such as CUSHING preferred aggres-
sive surgical approaches. They recognized the neces-
2 Material and methods
From February 1992 to December 1994,148 patients
with craniocerebral missile wounds were treated
surgically during the Southeastern Anatolia con-
flict. The patients were evaluated initially by the par-
amedics and transferred to semimobile surgery hos-
pital. At this point, broad spectrum antibiotics and
dexamethasone were given intravenously. After re-
suscitation, stabilization, and more immediate life-
saving procedures, the patients were transferred by
© 1997 by Walter de Gruyter & Co. Berlin • New York
178
Table
I. Causes of infection
G6nttl et al., Penetrating craniocerebral injuries
Organism No. of cases Total
With CSF fistula Without CSF fistula
Staphylococcus aureus 3 2 5
Staphylococcus coagulase (-) 1 - 1
Enterobacter species 1 - 1
Klebsiella pneumonia 1 - 1
Serratia marceinces 1 - 1
Table
lI. Infections types
Type of infection No. of cases With CSF fistulas Without CSF fistulas
Subdural empyema 2 Otorrhea -
Meningitis 2 Rhinorrhea -
Cerebral abscess 2 Rhinorrhea-Otorrhea 2
Purulent ventriculitis 1 Rhinorrhea -
the helicopter ambulance to the center where the
neurosurgical management was carried out. The
mean time lapse before arrival was two hours. Most
of the patients were operated on and followed by
senior author.
The mean age of patients was 22 years (range 4-45
years). Ninety-eight patients (66 %) were injured by
shell fragments and 50 patients (34 %) by gunshot.
The wound was classified as penetrating in 81 cases
(56 %), tangential for 37 cases (24 %), and perforat-
ing for 30 cases (20 %).
All patients had preoperative roentgenograms of
the skull and 12 patients in stable condition had ce-
rebral CT scans. All patients underwent surgery, and
preoperatively visible and palpable intracranial
bone and metal fragments were removed in the first
operation. Necrotic cerebral tissue was debrided.
Bone and metal fragments in a deep or intraventric-
nlar location were not removed. The dura was closed
primarily in 7 cases (4 %), with pericranium or tem-
poral fascia in 110 cases (74 %), tensor fascia lata
graft in 29 cases (20.5 %), and cadaver dura graft in
2 cases (1.5 %). All patients received ceftriaxone
4 g/day for 14 days and/or nidazol 1 g/day intra-
venously for 10 days postoperatively In addition,
the patients received standard medication protocols
including antiedema, antiepileptic, and analgesic
agents. Postoperatively, all patients were followed
periodically by CT scan. If the CT scan did not show
signs of infection retained intracranial bone and
metal fragments were left in place.
3 Results
The frequency and causes of infection due to retained
bone and metal fragments are shown in table I. None
of the retained metal fragments migrated on follow
up CT scans. The organisms cultured from CSF or
pus of our 9 patients and the infection types are
shown in table II.
Eleven of 148 patients (7.3 %) had CSF fistulas;
2 from the scalp wound, 2 from orbit, 3 from the ear,
4 from the nose. 7 (63 %) of the 11 patients with CSF
fistula developed central nervous system infections
(Figure 1). 51 patients had retained bone and metal
fragments. Only 2 patients had developed CNS in-
fections (Figures 2 and 3). The infection rate in pa-
tients with retained bone or metal fragments was
4 % compared to 7 % in patients without retained
bone or metal fragments, not statistically significant.
Twelve patients died in this series. The death rate
was 8 %. Posterior fossa and brain stem injury had
high mortality rate (shown in table III). Correlations
between death rate and GCS are shown in table IV.
The patients were followed for a mean of 12 months,
ranging between 5 months to 18 months. No patients
Neurosurg. Rev. 20 (1997)
G6nfil et al., Penetrating craniocerebral injuries 179
Figure 1. Left orbito-cranial injury CT scan showing metal
fragment passing through the posterior to the left cerebel-
lar lobe.
Figure 2. Brain abscess in the right parietal region in the
CT scan. A Bone fragment can be seen at the right para-
sagittal location.
Figure 3. CT scan showing multiple abscess in the left pa-
rietal lobe. A bone fragment is seen near the posterior horn
of lateral ventricle.
were lost during the follow-up period. All patients
were hospitalized for a long time at their first admis-
sion because of rehabilitation procedures. Only some
patients were rehospitalized secondarily for a recon-
structive surgery such as a cranioplasty. No patients
developed seizure disorders except the cases with
low plasma levels of phenytoin. No correlation was
found between bone fragments and seizure dis-
orders. Brain abscesses occurred usually in the sev-
enth week.
4 Discussion
Postdebridement infection is one of the most signif-
icant secondary causes of morbidity and mortality in
penetrating brain wounds [1, 2, 4, 6, 9, t0, 14]. Late
infections are a more common problem than menin-
gitis due to early operation and administration of
strong antibiotics. Antimicrobial agents provide
prophylaxis against invasion of bacteria, but they are
not effective in and around necrotic tissue, clot, and
foreign bodies [6, 7].
The surgeons of the periods of both World War II
and Korea War were less aggressive in their ap-
proach to the cases, and the results were higher in-
fection rates and increased mortality. Some of the
surgeons recorded that these infections were ob-
served especially when in-driven bone fragments re-
mained after the initial debridement. At reexplora-
tion, residual necrotic tissue or located pus was dis-
covered in close approximation of the retained frag-
ments in a large number of cases and culture of this
debris was frequently positive [3, 8, 9]. This experi-
ence advocate aggressive initial debridement.
According to MEIROWSKY [9], 75 % of all fragments
are probably sterile immediately after wounding
within brain, but there may be bacterial invasion
when neurosurgical procedures are delayed.
During World War II, MALTBY [3] reported that
there was no correlation between abscess formation
and the presence of retained bone fragments. He de-
scribed 17 cases complicated by abscess formations,
only three of which had retained bone fragments.
This was also supported by the experimental studies
of PITLYK [12], et al. In their research on bone and
debris implanted in canine cortex they found that
when sterile or contaminated bone alone was im-
planted the ratio of abscess development was 8 %
and 4 %, respectively. However the combination of
either with scalp or hair increased the rate to nearly
70 %.
In the Vietnam head injury study, 1221 patients were
reviewed 5 years after wounding. A total of 37 brain
abscess were discovered but, of these, only 11 had re-
tained bone fragments and none of these occurred
without at least other major risk factors, including
facio-orbital entry, CSF fistula, wound complica-
tions, prolonged coma, or multiple surgical proce-
dures [13].
In phase II of the study, 481 of these patients were
re-evaluated with CT and retained bone fragments
Neurosurg. Rev. 20 (1997)
180 GOntil et al., Penetrating craniocerebral injuries
Table III. Causes of
death
Cause of death No, of cases
Brain stem injury 3
Extensive brain damage 4
Central nervous system infection 2
Systemic infection 2
Pulmonary embolism 1
Table
IV. GCS and death rate (Correlation of death rate
with GCS)
GCS 4 and below 10 cases (83 %)
GCS 5 and above 2 cases (17 %)
were found in 23 % all with infection [11]. In the
113 patients with craniocerebral injuries examined
by BRANDVOLD et al. [3], they used less aggressive
approaches. Fourty-three of these cases were re-
evaluated after 5.9 years and 22 patients (51%)
were found to have retained intracranial bone frag-
ments on CT scan, but none had an infection of
central nervous system.
379 patients from Iran-Iraq War were reexamined
by AARABI who found chances of CNS infection to
be increased in those with retained bone frag-
ments. In their study there was a high incidence of
rhinorrhea and otorrhea in cases with CSF fistula
at the injury site. AARABI noted that in such pa-
tients the chance of developing an infection was in-
creased 20 times. AARAm [1, 2] suggested that re-
tained bone fragments increased the rate of infec-
tions of the central nervous system, although the
correlation did not show statistical significance. He
noted that although the age of wound at treatment
was increased to a few hours by helicopter evacu-
ation in Vietnam experience, the infection rate
(5.9 %) was higher than that rate in Iran-Iraq War
(4.7 %) where wounds were first treated after
49 hours. He concluded that wound age did not in-
crease the infection rate. It is emphasized that
even when the wound age is high, adequate de-
bridement of infected and contamined necrotic
tissues, dura plasty and antibiotic therapy de-
crease the infection rate [1].
In our series, 9 (6 %) of 148 patients developed in-
fection during postoperative period. Among these,
4 revealed intracerebral abscess, 2 subdural em-
pyema, 2 meningitis, and 1 purulent ventriculitis.
Seven patients had a CSF fistula (otorrhea and rhi-
norrhea), and 2 patients had intracranial bone
fragments without CSF fistula. The 2 patients who
had an intracranial bone fragment had intracere-
bral abscesses without CSF fistula. The infection
rate was increased in patients with CSF fistula [1,
10].
CAREY et al. [4] noticed contamination in 45 % of
cases with retained bone fragments and reported 2
minor and major complications, and one death after
secondary debridement in 103 patients. MEIROWS-
KY [9]
reported increasing neurological deficit in 4
(3.4 %) of 116 secondarily operated patients, These
reports show that secondary surgery to remove re-
tained fragments has the risk of increasing the neu-
rological deficit.
Metal fragments are, in general, sterile due to the
temperature of the metal. Most often, other infect-
ing materials and small bone fragments are trans-
ported into deep sites by metal fragments [4, 6]. In
13 our cases with deep injury, small metal fragments
were found, but in all these patients no infection was
determined.
MEIROWSKY et al. [9] found CSF fistulas in 101 of the
1113 patients with craniocerebral missile wounds
caused by bullets in the Vietnam War. Fifty (49.5 %)
of these patients developed infection. In the 1032
cases who had no CSF fistula only 47 (4.6 %) cases
revealed infection. Development of an abscess
around a bone fragment occurs within 6 to 8 weeks.
In our series, the cases of intracerebral abscess who
had an intracranial retained bone fragment showed
abscess formation on CT scan between the 6 th and
8 th weeks.
At the American University of Beirut Medical Cen-
ter TAHA et al. [16] treated 32 patients without deep
injury or intracranial hematoma and with a Glasgow
score higher than 10 out of 600 patients with pene-
trating craniocerebral missile wound using only pri-
mary wound debridement and closure. Only in a pa-
tient with CSF fistula and deeply located retained
bone fragment did brain abscess occur within 20
days after trauma. No complication was observed in
the other patients. The microorganisms which reach
deep sites of the brain due to on-going CSF fistulas
make management difficult. Despite use of antibio-
tics, contamination can not be prevented. The CSF
fistula is considered to be a maj or source of infection
[1, 10].
Neurosurg. Rev. 20(1997)
GGnfil et al., Penetrating craniocerebral injuries 181
Some authors suggest that asymptomatic intracra-
nial retained bone fragments should not be removed,
because reexploration can increase neurological def-
icits [3,11]. We conclude that the probability of brain
abscess due to retained bone fragments is low com-
pared to probability of increasing neurological defi-
cits. Our results are similar to those of previously re-
ported series. In this clinical study, we recommend
conservative approaches for clinically and radiologi-
cally asymptomatic intracranial retained bone and
metal fragments and periodic follow-up by CT scan.
MRI was not performed in the patients with metal
fragments, although MRI is the best imaging tech-
nique for the management of neurosurgical treat-
ment.
5 Conclusion
Our experience indicates that the following steps are
required in the surgical management of penetrating
craniocerebral missile wounds:
1. Debridement of necrotic tissues with hemostasis
and removal of foreign bodies without deep location.
2. When the CSF leakage occurs after dural repair
with primary material or grafting, lumbar puncture,
or lumbar drainage, or reoperation must be per-
formed to prevent CSF fistula.
3. If retained bone and metal fragments show no signs
of infection, cases should be periodically followed by
CT-scan. When there is evidence of infection, re-
operation should be performed.
References
[1] AARABI B:
Causes of infections in penetrating head
wounds in the Iran-Iraq War. Neurosurg 25 (1989)
923-926
[2] AARABI B: Comparative study of bacteriological con-
tamination between primary and secondary explora-
tion of missile head wounds. Neurosurg 20 (1987)
610- 616
[3] BRANOVOLD B, L LEVI, M FeINSOD, ED GEORGE:
Penetrating craniocerebral injuries in the Israeli in-
volvement in the Lebanese conflict. 1982-1985. J Neu-
rosurg 72 (1990) 15-21
[4] CAREY M, H YOUNG, JL MATHIS, J FORSYTE: A bac-
teriological study of craniocerebral missile wounds
from Vietnam. J Neurosurg 34 (1971) 145-154
[5] CLARK WC, MS MUHLBAUER, CB WATRIDGE, MW
RA'e. Analysis of 76 civilian craniocerebral gunshot
wounds. J Neurosurg 65 (1986) 9-14
[6] HAGAN R: Early complications following penetrating
wounds of the brain. J Neurosurg 34 (1971) 132-141
[7] HAMMON WM: Analysis of 2187 consecutive pene-
trating wounds of the brain from Vietnam. J Neuro-
surg 34 (1971) 127-131
• [8] HAMMON WM: Retained intracranial bone frag-
ments: Analysis of 42 patients. J Neurosurg 34 (1971)
142-144
[9] MEIROWSKY AM: Secondary removal of retained
bone fragments in missile wounds of the brain. J Neu-
rosurg 57 (1982) 617-621
[10] MEIROWSKY AM, WF CAVENESS, JD DILLON, BL
RISH, JP MOHR, JP KISTLER, GH WEISS: Cerebrospi-
nal fluid fistulas complicating missile wounds of the
brain. J Neurosurg 54 (1981) 44-48
[11]
MYERS PW, J BROPHY, AM SALAZAR: Retained bone
fragments after penetrating brain wounds: long-term
follow-up in Vietnam veterans. J Neurosurg 70 (1989)
319
[12] PITLYK P J, S TOLCHIN, W STEWART. The experimental
significance of retained intracranial bone fragments.
J Neurosurg 33 (1970) 19-24
[13] RISI~ BL, WF CAVENESS, JD DILLON: Analysis of
brain abscess after penetrating craniocerebral inju-
ries in Vietnam. J Neurosurg 9 (1981) 535-541
[14] RISH BL, JD DILLON, GH WEISS: Mortality following
penetrating craniocerebral injuries. J Neurosurg 59
(1983) 775-780
[15] SICCARDI D, R
CAVALIERE,
A PAU, F LUBINU, S TUR-
TAS, GL VIALE: Penetrating craniocerebral missile in-
juries in civilians: A retrospective analysis of 314
cases. Surg Neuro135 (1991) 455-46
[16] TAHA JM, MI
SABA,
JA BROWN: Missile injuries to
the brain treated by simple wound closure: Results of
a protocol during the Lebanese Conflict. J Neurosurg
29 (1991) 380-384
Submitted January 25, 1996.
Accepted July 17, 1996.
Engin GGnttl, MD.
Assistant Professor
Department of Neurosurgery
Faculty of Medicine
GATA
Ankara
Turkey
Neurosurg. Re,z 20 (1997)