364

Antibiotic Prophylaxis After Basilar Skull Fractures: A Meta-Analysis

Tibisay Villalobos, Carlos Arango, Paul Kubilis, Division of Pediatric Infectious Disease/Immunology, University of
Florida Health Science Center, Jacksonville, Nemours Childrens Clinic
and Mobeen Rathore
and Wolfson Childrens Hospital, Jacksonville, Florida
Antibiotic prophylaxis after basilar skull fractures remains controversial. Previous studies have
not clearly delineated the utility of prophylactic antibiotics in this setting. We undertook this
study to determine if antibiotic prophylaxis after basilar skull fractures prevented meningitis. We
performed a formal systematic review of previously published studies after a computerized search
with use of the MEDLINE data base (19701996). Fourteen studies were identied, and 12 studies
met the criteria for inclusion. Study design and quality were assessed by two independent investiga-
tors with use of a predetermined protocol. A total of 1,241 patients with basilar skull fractures were
included; 719 patients received antibiotics, and 522 patients did not receive antibiotics. Overall
results suggest that antibiotic prophylaxis did not prevent meningitis among patients with basilar
skull fractures (odds ratio [OR] 1.15; 95% condence interval [CI] 0.681.94; P .678).
Patients with basilar skull fractures and cerebrospinal uid leakage were analyzed separately
(OR 1.34; 95% CI 0.752.41; P .358), as were children (OR 1.04; 95% CI 0.0714.90;
P 1.000). Antibiotic prophylaxis after basilar skull fractures does not appear to decrease the risk
of meningitis.
Basilar skull fractures account for 19% (range, 3%24%) vents the development of bacterial meningitis in patients with
basilar skull fractures. of all skull fractures [1, 2]. Bacterial meningitis, one of the most
serious complications of basilar skull fracture, is attributed to
communication of the subarachnoid space with contaminated
extracranial cavities of the paranasal sinuses, nasopharynx, or Materials and Methods
middle ear. The reported incidence of meningitis after basilar
We searched the medical literature for published studies con-
skull fracture varies from 9.2% to 17.8% and can be as high
cerning the use of prophylactic antibiotics for prevention of
as 50% if CSF leakage is present [35].
meningitis after basilar skull fracture. Computerized searches
It has been suggested that meningitis after basilar skull frac-
using MEDLINE between 1970 and 1996 were performed inde-
ture might be prevented with the use of prophylactic antibiotics
pendently by two different investigators and a professional
[6, 7]; however, such antibiotic prophylaxis remains controver-
medical librarian using the following key words: basilar skull
sial. Some experts recommend against the use of prophylactic
fracture, meningitis, and antibiotics. Those original articles
antibiotics [1, 2, 4, 8, 9], whereas others recommend routine
and abstracts that indicated antibiotic use for prevention of
antibiotic prophylaxis for a period ranging from 314 days or
meningitis among patients with basilar skull fractures were
until 1 week after CSF leakage has resolved [6, 10, 11]. Most
evaluated. Bibliographies of all relevant original articles and
studies performed to date are retrospective and do not show a
review articles were examined to identify any additional stud-
difference in the development of meningitis among patients
ies. All reports were reviewed by two separate investigators
who receive antibiotic prophylaxis vs. those who do not. In
who were blinded to one anothers review strategies. The qual-
addition, because of the small samples there is insufcient
ity of each study was evaluated according to a previously pub-
power to demonstrate statistically signicant differences be-
lished instrument for assessing validity of primary studies and
tween patients who did and patients who did not receive pro-
assigned a score ranging from 0 to 12 [12]. A third investigator
phylactic antibiotics.
served as a referee for the rst two investigators in the event
Meta-analysis can generate a more precise estimate of treat-
of any signicant disagreements.
ment effect by combining results across many studies and thus
For each article, a two-by-two contingency table was created
increasing the statistical power and precision. We applied meta-
to summarize the number of patients reported as having menin-
analytic techniques to determine if antibiotic prophylaxis pre-
gitis or not having meningitis in each group, and those who
did not receive antibiotics vs. those who received antibiotics.
The measure of effect was taken as the change in odds of
infection for subjects not receiving prophylactic antibiotics rel- Received 8 October 1997; revised 31 March 1998.
Reprints or correspondence: Dr. Mobeen H. Rathore, 653-1 West 8th Street,
ative to those subjects receiving prophylactic antibiotics. Exact
Jacksonville, Florida 32209.
ORs for meningitis risk (i.e., the odds of developing meningitis
Clinical Infectious Diseases 1998; 27:3649
in untreated patients relative to patients treated with prophylac-
1998 by the Infectious Diseases Society of America. All rights reserved.
10584838/98/27020018$03.00 tic antibiotics) and 95% CIs were estimated for each individual
/ 9C52$$AU34 07-09-98 19:53:09 cida UC: CID

b
y

g
u
e
s
t

o
n

J
u
l
y

7
,

2
0
1
4
h
t
t
p
:
/
/
c
i
d
.
o
x
f
o
r
d
j
o
u
r
n
a
l
s
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

365 CID 1998; 27 (August) Basilar Skull Fractures
study by using methods developed initially by Mehta et al. was started within the rst 72 hours of hospitalization and
continued for 3 days to at least 1 week after CSF leakage had [13] and available from the EGRET epidemiological statistics
software (version 1; Cytel Software, Cambridge, MA) [13]. An resolved. Information about the specic length of follow-up
was insufcient. Table 1 summarizes the description of the exact common OR test was also performed using EGRET to
determine if corresponding ORs from individual studies dif- studies. A total of 1,241 patients with diagnoses of basilar
skull fractures were analyzed. Of these, 719 patients received fered signicantly among each other. If the common OR test
was not signicant, EGRET was used to estimate an exact antibiotics and 522 did not receive antibiotics. None of the
individual study ORs for meningitis risk (no prophylactic anti- common OR with 95% CI. The OR for meningitis risk that
could be detected with a power of 80% at a signicance level of biotics vs. prophylactic antibiotics) differed signicantly from
1. An OR could not be estimated for the study by Einhorn and .05, given the overall meta-analysis sample size and individual
study sample fractions and meningitis rates, was also computed Mizrahi [14] because of the absence of cases of meningitis in
either group. ORs estimated for the other 11 individual studies using the statistical power calculation module (SIZ module) of
EGRET. Because CSF leakage has been associated with a ranged from zero to innity. There were no signicant differ-
ences among the ORs estimated for individual studies greater risk of meningitis, an additional analysis using similar
methods was done for this subgroup of patients. (P .443). The common OR estimated across all studies
(OR 1.15; CI 0.681.94; P .678), although indicating To assess the inuence of individual studies on the estimated
common OR, a sensitivity analysis was carried out by eliminat- decreased meningitis risk among patients treated with prophy-
lactic antibiotics, did not differ signicantly from 1. Retrospec- ing each of the 12 studies one at a time and reestimating the
common OR and 95% CIs for the remaining 11 studies and, tive analysis of statistical power indicated that, given the over-
all sample size and individual study sample fractions and nally, comparing the OR and CI for the 11 studies to the
common OR estimate for the 12 studies analyzed together. In reported meningitis rates, an OR of 1.62 could be detected
with 80% power at a signicance level of .05. Results for a similar manner, common ORs and 95% CIs were estimated
separately for the two prospective studies and the 10 retrospec- individual study analyses and combined study analyses are
summarized in table 2 and gure 1. tive studies.
A sensitivity analysis was performed to assess the inuence
of individual studies on the results of the meta-analysis by
Results
excluding each of the 12 studies one by one and reperforming
the meta-analysis on the remaining 11 studies. The common A total of 14 studies were identied. Two studies were ex-
cluded from the analysis because no specic data could be OR estimated for these 12 separate analyses with 11 studies
ranged from 0.99 to 1.31. As was the case with the common extracted. One of these two studies was published by Hoff and
Brewin [1], in a letter to the editor, and described 160 patients OR of the 12 studies analyzed together (OR 1.15), the com-
mon ORs estimated in the sensitivity analysis did not differ with basilar skull fractures and no CSF leakage, randomized
blindly to receive either one of two different prophylactic anti- signicantly from 1 (P .329.728). The exclusion of the
study by Eljamel [19], which had both the largest number of biotic regimens or no prophylactic antibiotics. No cases of
meningitis developed within the three groups, but no data about patients and cases of meningitis, had the greatest effect on the
precision of the common OR estimate, not surprisingly yielding the patients were available for analysis [1]. The other excluded
study was published by Leech and Paterson in 1973 [6] and the widest 11-study 95% CI among all intervals estimated in
the sensitivity analysis (OR 0.99; 95% CI 0.422.25). retrospectively compared the conservative and operative man-
agement of basilar skull fractures with CSF leakage. In addi- The common OR estimated separately for retrospective studies
was 1.17 (95% CI 0.682.01; P .706) and for the prospec- tion, the authors compared the incidence of meningitis associ-
ated with rhinorrhea and otorrhea. Data regarding the use of tive studies was 0.68 (95% CI 0.0113.77; P .187).
prophylactic antibiotics in the two groups of patients could not
be adequately determined.
Antibiotic Prophylaxis in Patients with Basilar Skull Fractures
Of the 12 studies selected for meta-analysis, nine were retro-
and CSF Leakage
spective [11, 1421], two were prospective and randomized
[22, 23], and one was a combined prospective-retrospective Data from nine studies that included patients with CSF leakage
were available for comparison and analysis [11, 14, 15, 17, 20 study [2]. Among the retrospective studies different antibiotic
regimens were used, including penicillins (penicillin G, ampi- 23]. These nine studies described 547 patients with CSF leakage,
297 of whom received prophylactic antibiotics (29 developed cillin, or semisynthetic penicillin alone or in combination),
rst-generation cephalosporins, chloramphenicol, sulfon- meningitis) and 250 of whom did not (34 developed meningitis).
When studies were analyzed individually, none of the estimated amides, and gentamicin in combination with penicillin. In the
more recent studies, third-generation cephalosporins were used. ORs differed signicantly from 1. The P value for the common
OR test (P .278) indicated that there were no signicant differ- In the two prospective studies, ceftriaxone, ampicillin/sulfadia-
zine, and penicillin were used. In all studies, antibiotic therapy ences among the ORs estimated for individual studies. The com-
/ 9C52$$AU34 07-09-98 19:53:09 cida UC: CID

b
y

g
u
e
s
t

o
n

J
u
l
y

7
,

2
0
1
4
h
t
t
p
:
/
/
c
i
d
.
o
x
f
o
r
d
j
o
u
r
n
a
l
s
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

366 Villalobos et al. CID 1998; 27 (August)
Table 1. Description of studies of basilar skull fractures and use of prophylactic antibiotics.
No. of No. of patients
patients with No. of without No. of
Study no. No. of prophylactic patients with prophylactic patients with
[reference] Type of study patients antibiotics meningitis antibiotics meningitis
1 [2] Prospective/retrospective 104 54 2 50 0
2 [22] Prospective/randomized 43 30 0 13 1
3 [23] Prospective/double-blind 52 26 0 26 1
4 [11] Retrospective 58 41 1 17 2
5 [15] Retrospective 161 78 2 83 2
6 [16] Retrospective 42 28 2 14 1
7 [17] Retrospective 347 251 4 96 2
8 [18] Retrospective-pediatric 48 23 1 25 1
9 [19] Retrospective 215 106 20 109 25
10 [20] Retrospective-pediatric 37 20 1 17 1
11 [21] Retrospective 88 48 5 40 1
12 [14] Retrospective-pediatric 46 14 0 32 0
Total 1,241 719 38 522 37
mon OR estimated across all studies (OR 1.34; 95% CI an OR of 1.78 could be detected with 80% power at a
signicance level of .05. 0.752.41; P .358), although indicating lower meningitis risk
among patients who received antibiotics, did not differ signi- In addition, the risk of meningitis among patients with either
CSF rhinorrhea or CSF otorrhea was evaluated. Data from six cantly from 1 (table 3; gure 2).
Eljamel [19] reported 6.6% and 9.12% incidences of men- studies were available for meta-analysis. A total of 179 patients
were analyzed; 70 patients had CSF rhinorrhea, and 109 pa- ingitis during the rst week after basilar skull fractures
among patients receiving and not receiving prophylactic an- tients had CSF otorrhea. ORs could not be estimated for two
studies because there were no cases of meningitis reported [14, tibiotics, respectively. He also described the largest number
of patients with unrepaired CSF stulae. There was no statis- 18]. Among the patients with CSF rhinorrhea, four patients
developed meningitis, and among the patients with CSF otor- tically signicant difference in the risk of meningitis for
either group (OR 1.28; 95% CI 0.632.63; P .57). rhea, three developed meningitis. The site of CSF leakage (rhi-
norrhea vs. otorrhea) was not a predictor of risk for meningitis. Overall results were relatively unaffected by exclusion of
this study in the sensitivity analysis. Analysis of statistical The common OR did not differ signicantly from 1 (P .261).
There were no signicant differences among the ORs estimated power indicated that, given the overall sample size, individ-
ual study sample fractions, and reported meningitis rates, for the individual studies. The common OR estimated across
Table 2. Cases of meningitis after basilar skull fractures with and without antibiotic prophylaxis.
No. of cases of No. of cases of
meningitis without meningitis with
Authors [reference] antibiotics (%) antibiotics (%) OR* 95% CI P value

Ignelzi and VanderArk [2] 0 . . . 2 (3.7) 0 05.74 .534

Demetriades et al. [22] 1 (7.7) 0 . . . 0.06 .605
Klastersky et al. [23] 1 (3.9) 0 . . . 0.03 1.000
MacGee et al. [11] 2 (11.8) 1 (2.4) 5.16 0.25322.28 .406
Dagi et al. [15] 2 (2.4) 2 (2.6) 0.94 0.0713.25 1.000
Zrebeet and Huang [16] 1 (7.1) 2 (7.1) 1 0.0220.91 1.000
Frazee et al. [17] 2 (2.1) 4 (1.6) 1.31 0.129.33 1.000
Ash et al. [18] 1 (4.0) 1 (4.4) 0.92 0.0175.19 1.000
Eljamel [19] 25 (22.9) 20 (18.9) 1.28 0.632.63 .572
McGuirt and Stool [20] 1 (5.9) 1 (5.0) 1.18 0.0198.06 1.000
Clemenza et al. [21] 1 (2.5) 5 (10.4) 0.22 02.13 .298
Einhorn and Mizrahi [14] 0 . . . 0 . . . . . . 000 1.000
* Common OR 1.15; 95% CI 0.681.94; P .678.

P value for common OR .233.

/ 9C52$$AU34 07-09-98 19:53:09 cida UC: CID

b
y

g
u
e
s
t

o
n

J
u
l
y

7
,

2
0
1
4
h
t
t
p
:
/
/
c
i
d
.
o
x
f
o
r
d
j
o
u
r
n
a
l
s
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

367 CID 1998; 27 (August) Basilar Skull Fractures
them had CSF leakage. Two of the four patients who developed
meningitis received prophylactic antibiotics (table 4). Einhorn
and Mizrahi [14] did not report any cases of meningitis among
46 children with basilar skull fractures; 34 patients received
prophylactic antibiotics, and 32 did not, and only seven children
had CSF leakage. Although the sample was small, antibiotic
prophylaxis after basilar skull fractures in children did not
appear to prevent meningitis (OR 1.04; 95% CI 0.07
14.90; P 1.000).
Discussion
Antibiotic prophylaxis after basilar skull fracture, with or
without CSF leakage, remains controversial because the propo-
nents of prophylactic antibiotic use believe that because CSF
is exposed to potentially pathogenic organisms in the upper
respiratory tract (nose or ear), patients with basilar skull frac-
tures are at increased risk for meningitis and would thus benet
from antibiotic prophylaxis. The opponents argue that antibiotic
prophylaxis contributes to the development of potentially resis-
Figure 1. Inuence of antibiotic prophylaxis on meningitis risk tant organisms and more serious infection, and that ultimately
after basilar skull fractures: estimated ORs (points) and exact 95%
antibiotic prophylaxis does not decrease the risk of meningitis.
CIs (no antibiotics vs. antibiotics). OR lower 95% condence limit
Strong data indicating benet or lack of benet of antibiotic
.001 (arrows pointing to the left). OR upper 95% condence limit
prophylaxis after basilar skull fracture are not available, thus
1,000 (arrows pointing to the right). (An OR and CI could not be
further complicating the issue.
estimated for the Einhorn and Mizrahi study [14] because there were
no meningitis cases observed in that study.) Rathore [9], in a review of the articles published between
1970 and 1989, found 848 cases of basilar skull fractures (519
received antibiotic prophylaxis and 8% developed meningitis)
and concluded that antibiotics are not useful in preventing all the studies (OR 1.74; 95% CI 0.2613.36; P .772)
did not differ signicantly from 1. meningitis after basilar skull fracture [9].
The overall results of this meta-analysis of 1,241 patients
among 12 published studies suggest that antibiotic prophylaxis
Basilar Skull Fractures Among Children
does not decrease the risk of meningitis. The common OR of
1.15, although indicating an increased meningitis risk among Three studies were performed exclusively in children and
included 131 patients with basilar skull fractures [14, 18, 21]. patients not treated with antibiotics, did not reach statistical
signicance. A similar result was evident for patients with Fifty-seven children received prophylactic antibiotics, and 74
did not. Four patients (3%) developed meningitis, and all of CSF leakage. At 80% power, the meta-analysis could detect a
Table 3. Cases of meningitis after basilar skull fractures and CSF leakage with and without antibiotic prophylaxis.
No. of cases of No. of cases of
meningitis without meningitis with
Author [reference] antibiotics (%) antibiotics (%) OR* 95% CI P value

Demetriades et al. [22] 1 (11.1) 0 . . . 0.05 .643

Klastersky et al. [23] 1 (3.8) 0 . . . 0.03 1.000
MacGee et al. [11] 2 (11.8) 1 (2.4) 5.16 0.25322.28 .406
Dagi et al. [15] 2 (10.0) 0 . . . 0.12 1.000
Zrebeet and Huang [16] 1 (12.5) 2 (9.5) 1.34 0.022980 .406
Eljamel [19] 25 (22.9) 20 (18.9) 1.28 0.632.63 .572
McGuirt and Stool [20] 1 (5.9) 1 (5.0) 1.18 0.0198.06 1.000
Clemenza et al. [21] 1 (2.5) 5 (10.4) 0.22 0.002.13 .027
Einhorn and Mizrahi [14] 0 . . . 0 . . . . . . 0.00 1.000
* Common OR 1.34; 95% CI .75241; P .358.

P value for common OR .278.

/ 9C52$$AU34 07-09-98 19:53:09 cida UC: CID

b
y

g
u
e
s
t

o
n

J
u
l
y

7
,

2
0
1
4
h
t
t
p
:
/
/
c
i
d
.
o
x
f
o
r
d
j
o
u
r
n
a
l
s
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

368 Villalobos et al. CID 1998; 27 (August)
skull fractures, 43 of whom had basilar skull fractures. The
only patient who developed meningitis had a basilar skull frac-
ture and CSF otorrhea and did not receive prophylactic antibiot-
ics. The other study, by Klastersky et al. [23], compared pa-
tients with CSF rhinorrhea to those with CSF otorrhea. One
patient who was not receiving antibiotic prophylaxis developed
bacterial meningitis [22]. The statistical analysis performed
separately, including these two studies, showed that antibiotic
prophylaxis after basilar skull fracture does not decrease the
risk of bacterial meningitis; however, this result was not statisti-
cally signicant.
Several authors [10, 24, 25] have reported higher incidences
of meningitis with CSF rhinorrhea vs. CSF otorrhea. This nd-
ing has been attributed to the direct communication between
the cribriform plate and other basilar structures within the sinus
cavity. The present meta-analysis showed no statistically sig-
nicant difference in the risk of meningitis with CSF rhinorrhea
Figure 2. Inuence of antibiotic prophylaxis on meningitis risk
vs. CSF otorrhea.
after basilar skull fracture with CSF leakage: estimated OR (points)
and exact 95% CIs (no antibiotics vs. antibiotics). OR lower 95% Children represent a special category. The incidence of ba-
condence limit .001 (arrows pointing to the left). OR upper
silar skull fractures is lower than that for the adult population.
95% condence limit 1,000 (arrows pointing to the right). (An
Fracture of the skull is said to occur among 6%14% of chil-
OR and CI could not be estimated for the Einhorn and Mizrahi
dren with head trauma [2426]. Second, the adult-to-child ratio
study [14] because there were no meningitis cases observed in that
for traumatic CSF stulae is 10:1. This lower incidence is
study.)
thought to be related to the exibility of childrens skull bases
and to the underdevelopment of the ethmoid, frontal, and mas-
minimum of 62% increase in the odds of developing meningitis toid air cells. Third, CSF otorrhea is more common among
among untreated patients relative to patients treated with antibi- children than is CSF rhinorrhea, perhaps because of the higher
otics. Thus, if a true meningitis rate of 8% is assumed for incidence of temporal-bone fractures in children [2126]. Lau
untreated patients, a decrease in the true rate to 5% for patients and Kenna [8] reported an incidence of posttraumatic meningi-
treated with antibiotics could be detected with adequate statisti- tis of 0.38% in a 15-year review of children hospitalized with
cal power. head injuries [8]. Liu-Shindo et al. [27], in a series of 62
The incidence of bacterial meningitis is increased with ba- children with basilar skull fractures, reported two cases of men-
silar skull fractures and CSF leakage [3, 4, 24]. Most CSF ingitis, both associated with fractures of the paranasal sinuses.
stulae will resolve during the rst 710 days after a fracture; In our review, only four cases of meningitis were reported
surgical intervention is indicated if CSF leakage persists for among the studies of children, for an incidence of 3%. Although
several weeks [24, 25]. Surgery is seldom indicated during the the number of children is small, the use of antibiotic prophy-
acute phase, given that there is no difference in outcome with laxis in children with CSF leakage after basilar skull fracture
surgical repair during the rst week [4]. Antibiotics are usually did not prevent the development of meningitis and thus should
selected on the basis of the nasopharyngeal ora being the not be recommended on a routine basis.
source of the infection, and selection varies widely among
different institutions. The length of overall follow-up is im-
portant to assess the long-term outcome for patients who may
Table 4. Data from three studies of basilar skull fractures and men-
develop meningitis later after having received antibiotic pro-
ingitis among children with and without antibiotic prophylaxis.
phylaxis. Studies concerning follow-up of patients are rare. The
No. of patients study by Eljamel [19] specically evaluated the issue of follow-
No. of patients without
up among these patients for 6 years and found that overall
with antibiotics/ antibiotics/no. of
survival rate free of meningitis did not differ signicantly be-
No. of no. of patients patients with
tween the two groups. In addition, in another study [18], a
Author [reference] patients with meningitis meningitis
child developed pneumococcal meningitis 122 days after injury
Ash et al.* [18] 48 23/1 25/1 and was treated with prophylactic antibiotics for 1 week after
McGuirt and Stool

[20] 37 20/1 17/1

clinical resolution of CSF leakage.
Einhorn and Mizrahi [14] 46 14/0 32/0
There were only two controlled randomized studies that
compared prophylactic antibiotic use after basilar skull fracture
* Patients with temporal bone fractures and CSF leakage.

The two patients with meningitis had CSF leakage. [22, 23]. Demetriades et al. [22] analyzed 157 patients with
/ 9C52$$AU34 07-09-98 19:53:09 cida UC: CID

b
y

g
u
e
s
t

o
n

J
u
l
y

7
,

2
0
1
4
h
t
t
p
:
/
/
c
i
d
.
o
x
f
o
r
d
j
o
u
r
n
a
l
s
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

369 CID 1998; 27 (August) Basilar Skull Fractures
13. Mehta CR, Patel NR, Gray R. Computing an exact interval for the common
Conclusion
odds ratio in several 2 1 2 contingency tables. JASA 1986; 80:969
A meta-analysis was applied to summarize the results of 12 79.
14. Einhorn A, Mizrahi EM. Basilar skull fractures in children. Am J Dis
published studies concerning antibiotic prophylaxis for bacte-
Child 1978; 132:11214.
rial meningitis after basilar skull fractures. The results do not
15. Dagi FT, Meyer FD, Poletti CA. The incidence and prevention of meningi-
support the use of prophylactic antibiotics to decrease the risk
tis after basilar skull fracture. Am J Emerg Med 1983; 3:2958.
of bacterial meningitis in patients with basilar skull fractures.
16. Zrebeet HAK, Huang PS. Prophylactic antibiotics in the treatment of
fractures at the base of the skull. Del Med J 1986; 58:7418.
17. Frazee RC, Mucha P, Farnell MC, Ebersold MJ. Meningitis after basilar
References
skull fracture: does antibiotic prophylaxis help? Postgrad Med 1988;
85:26774.
1. Hoff JT, Brewin A. Antibiotics for basilar skull fractures [letter]. J Neuro-
18. Ash GJ, Peter J, Bass DH. Antimicrobial prophylaxis for fractured base
surg 1976; 44:649.
of skull in children. Brain Inj 1992; 6:5217.
2. Ignelzi RJ, VanderArk GD. Analysis of the treatment of basilar skull
19. Eljamel MS. Antibiotic prophylaxis in unrepaired CSF stulae. Br J Neuro-
fractures with and without antibiotics. J Neurosurg 1975; 43:7216.
surg 1993; 7:5016.
3. Marion DW. Complications of head injury and their therapy. Neurosurg
20. McGuirt WF, Stool SE. Cerebrospinal uid stula: the identication and
Clin North Am 1991; 2:41123.
management in pediatric temporal bone fractures. Laryngoscope 1995;
4. Kaufman BA, Tunkel AR, Pryor JC, Dacey RE. Meningitis in the neurosur-
105:35964.
gical patient. Infect Dis Clin North Am 1990; 4:677701.
21. Clemenza JW, Kaltman SI, Diamon DL. Craniofacial trauma and cerebro-
5. Helling TS, Evans LL, Fowler DL, et al. Infectious complications in
spinal uid leakage: a retrospective clinical study. J Oral Maxillofac
patients with severe head injury. J Trauma 1988; 28:15757.
Surg 1995; 53:10047.
6. Leech JP, Paterson A. Conservative and operative management for cere-
22. Demetriades D, Charalambides D, Lakhoon I, Pantnowitz D. Role of
brospinal uid leakage after closed head injury. Lancet 1973; 1:
prophylactic antibiotics in open and basilar fractures of the skull: a
10136.
randomized study. Injury 1992; 23:277380.
7. Brawley BW, Kelly WA. Treatment of basilar skull fracture with and
23. Klastersky UJ, Sadegli M, Brihaye J. Antimicrobial prophylaxis in patients without CSF stulae. J Neurosurg 1967; 26:5760.
with rhinorrhea or otorrhea: a double-blind study. Surg Neurol 1976; 8: 8. Lau YL, Kenna AP. Post-traumatic meningitis in children. Injury 1986;
1114. 17:4079.
24. Chandler JR. Traumatic cerebrospinal uid leakage. Otolaryngol Clin 9. Rathore MH. Do prophylactic antibiotics prevent meningitis after basilar
North Am 1983; 16:62332. skull fracture? Pediatr Infect Dis J 1991; 10:878.
25. Caldicott WJH, North JB, Simpson DA. Traumatic cerebrospinal uid 10. Leech P. Cerebrospinal uid leakage, dural stulae and meningitis after
stulas in children. J Neurosurg 1973; 38:19. basal skull fractures. Injury 1974; 6:1419.
26. Mealey J Jr. Skull fractures. In: Pediatric neurosurgery. Palisades, NY: 11. MacGee EE, Cauthern JC, Brackett CE. Meningitis following CSF stula.
Grune & Stratton, 1982:2929. J Neurosurg 1970; 33:3126.
27. Liu-Shindo M, Hawkins DB. Basilar skull fractures in children. Int J 12. Labbe KA, Detsky AS, ORourke K. Meta-analysis in clinical research.
Ann Intern Med 1987; 1097:22433. Pediatr Otorhinolaryngol 1989; 17:10917.
/ 9C52$$AU34 07-09-98 19:53:09 cida UC: CID

b
y

g
u
e
s
t

o
n

J
u
l
y

7
,

2
0
1
4
h
t
t
p
:
/
/
c
i
d
.
o
x
f
o
r
d
j
o
u
r
n
a
l
s
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m