clinical article

J Neurosurg Pediatr 17:595–601, 2016

Pediatric intracranial gunshot wounds: the Memphis

experience
Michael DeCuypere, MD, PhD,1 Michael S. Muhlbauer, MD,1–3 Frederick A. Boop, MD,1–3 and
Paul Klimo Jr., MD, MPH1–3
1
Department of Neurosurgery, University of Tennessee Health Science Center; 2Semmes-Murphey Neurologic and Spine
Institute; and 3Le Bonheur Neuroscience Institute, Le Bonheur Children’s Hospital, Memphis, Tennessee

Objective Penetrating brain injury in civilians is much less common than blunt brain injury but is more severe over-
all. Gunshot wounds (GSWs) cause high morbidity and mortality related to penetrating brain injury; however, there are
few reports on the management and outcome of intracranial GSWs in children. The goals of this study were to identify
clinical and radiological factors predictive for death in children and to externally validate a recently proposed pediatric
prognostic scale.
Methods The authors conducted a retrospective review of penetrating, isolated GSWs sustained in children whose
ages ranged from birth to 18 years and who were treated at 2 major metropolitan Level 1 trauma centers from 1996
through 2013. Several standard clinical, laboratory, and radiological factors were analyzed for their ability to predict
death in these patients. The authors then applied the St. Louis Scale for Pediatric Gunshot Wounds to the Head, a
scoring algorithm that was designed to provide rapid prognostic information for emergency management decisions.
The scale’s sensitivity, specificity, and positive and negative predictability were determined, with death as the primary
outcome.
Results Seventy-one children (57 male, 14 female) had a mean age of 14 years (range 19 months to 18 years). Over-
all mortality among these children was 47.9%, with 81% of survivors attaining a favorable clinical outcome (Glasgow Out-
come Scale score ≥ 4). A number of predictors of mortality were identified (all p < 0.05): 1) bilateral fixed pupils; 2) deep
nuclear injury; 3) transventricular projectile trajectory; 4) bihemispheric injury; 5) injury to ≥ 3 lobes; 6) systolic blood
pressure < 100 mm Hg; 7) anemia (hematocrit < 30%); 8) Glasgow Coma Scale score ≤ 5; and 9) a blood base deficit
< -5 mEq/L. Patient age, when converted to a categorical variable (0–9 or 10–18 years), was not predictive. Based on
data from the 71 patients in this study, the positive predictive value of the St. Louis scale in predicting death (score ≥ 5)
was 78%.
Conclusions This series of pediatric cranial GSWs underscores the importance of the initial clinical exam and CT
studies along with adequate resuscitation to make the appropriate management decision(s). Based on our population,
the St. Louis Scale seems to be more useful as a predictor of who will survive than who will succumb to their injury.
http://thejns.org/doi/abs/10.3171/2015.7.PEDS15285
Key Words pediatric intracranial gunshot wound; survival; traumatic brain injury; pediatric management; trauma

U
nfortunately, intracranial ballistic injuries due to be rising. For instance, from 2004 to 2008, the mortality
firearms, most commonly inflicted by low-velocity rate from intracranial GSWs (all ages) was 6.10 deaths per
handguns, are not uncommon events at many ma- 100,000 population (http://www.cdc.gov/injury/wisqars/
jor US metropolitan Level 1 trauma centers. They often index.html). In the years 2004–2010, this rate rose to
lead to severe neurological injury or death. Adult mortal- 6.34 deaths per 100,000 population. Unfortunately, the
ity rates from such injuries range between 50% and 90% same trend was noted for pediatric patients (1.31 deaths
in most series.1,6,9,10,12,13,15,20,27,31 Data from the Centers for per 100,000 population in 2004–2008 vs 1.42 deaths per
Disease Control and Prevention (CDC) suggest that deaths 100,000 population in 2004–2010). In fact, firearm injury
due to intracranial gunshot wounds (GSWs) in the US may was the fifth leading violent cause of hospitalization for

Abbreviations GCS = Glasgow Coma Scale; GOS = Glasgow Outcome Scale; GSW = gunshot wound; ICP = intracranial pressure; INR = international normalized
ratio; SBP = systolic blood pressure.
submitted May 8, 2015. accepted July 28, 2015.
include when citing Published online January 5, 2016; DOI: 10.3171/2015.7.PEDS15285.

©AANS, 2016 J Neurosurg Pediatr Volume 17 • May 2016 595

Unauthenticated | Downloaded 04/21/24 01:20 PM UTC

M. DeCuypere et al.

US children in 2013 (http://www.cdc.gov/injury/wisqars/ monitor (intraparenchymal sensor or external ventricular

index.html). drain) had been placed, the goal ICP was less than 20 cm
In the state of Tennessee, the annual death rate from H2O or mm Hg, utilizing any number of standard ICP-
intracranial GSWs (all ages) was 9.19 deaths per 100,000 lowering interventions. Concurrently, we attempted to
population from 2004 to 2010 (http://www.cdc.gov/ maintain the cerebral perfusion pressure above 40–60
injury/wisqars/index.html). The number of deaths over mm Hg, depending on the age of the child, by using in-
this time period was highest in Shelby County, where travenous fluids and vasopressors. Barbiturate coma was
561 people died from a GSW to the head. The annual induced only if all medical and surgical options had been
death rate for intracranial GSWs in pediatric patients in exhausted.
the state of Tennessee was 1.60 deaths per 100,000 popu- When surgery was performed, we followed principles
lation, and again the rate was highest in Shelby County consistent with what was published in the American As-
with 51 deaths over this 6-year period (http://www.cdc. sociation of Neurological Surgeons (AANS)/Congress of
gov/injury/wisqars/index.html). Neurological Surgeons (CNS) guidelines for the surgical
Pediatric GSWs present a challenge to pediatric neuro- management of penetrating brain injury.29 Surgery was
surgeons owing to their relative rarity and the consequent undertaken in all adequately resuscitated patients who we
paucity of literature. As suggested by some, caution should believed could survive their injury. It is our general phi-
be exercised when making clinical corollaries from the losophy that if the neurological exam and imaging stud-
literature on adults to the pediatric population.2,7,11 As is ies support meaningful survival or if there is a reasonable
the general rule for traumatic brain injury of any etiol- chance of this occurring, surgery should not be delayed by
ogy, children typically have lower overall mortality and gathering ICP data. Operative intervention ranged from
a greater propensity for neurological recovery than their wound debridement and small craniotomy to full decom-
adult counterparts, making it more difficult to prognosti- pressive craniectomy with duraplasty. Indications for de-
cate and administer appropriate intervention(s).2,7,11,16,19,22,24 compressive craniectomy included, but were not limited
In 2012, Bandt et al. proposed the St. Louis Scale for to, significant preoperative midline shift as a result of
Pediatric Gunshot Wounds to the Head.2 This tool, based cerebral edema and/or intracranial hematoma, intraopera-
on physical examination and imaging findings, is a scor- tive or anticipated postoperative brain swelling (preemp-
ing algorithm that rapidly provides prognostic information tive decompressive craniectomy), or intractable intracrani-
to clinicians for guidance in emergency management de- al hypertension occurring after primary debridement and
cisions. The purpose of the present study was 2-fold: 1) not responsive to maximum medical management.
to review our pediatric GSW experience in an attempt to
identify predictors of death, and 2) to externally validate Data Collection and Analysis
the St. Louis scale. Patient demographics, initial neurological examination
results and Glasgow Coma Scale (GCS) score, initial ICP
Methods (if applicable), vital signs, initial trauma laboratory values,
Study Questions and details of the hospital stay were obtained from the
chart review. Initial CT findings were analyzed for the fol-
We sought to answer 2 questions with our research: 1) lowing: 1) penetration of the deep nuclei (thalamus, cau-
What are the predictors of mortality in children who have date, putamen, and globus pallidus) and/or third ventricle
sustained a cranial penetrating (that is, through the skull, by bone or metallic fragments; 2) any shift of the midline
dura mater, and brain) GSW? 2) Can the St. Louis scale structures; 3) number of lobes penetrated (single lobe des-
accurately predict survival and death in our patient popu- ignation for infratentorial injury); 4) unilateral or bilateral
lation? hemispheric involvement; 5) intracranial compartments
involved (supratentorial, infratentorial, or both); and 6)
Patient Selection and Outcomes transventricular projectile trajectory. The St. Louis scale
This retrospective study was approved by the University score, as detailed in the original publication, was calcu-
of Tennessee Health Science Center Institutional Review lated retrospectively for each patient.2 In summary, the
Board. All pediatric patients presenting with gunshot in- score is composed of 9 items: 3 primary (3 points each),
juries to the head who received care in Memphis, Tennes- 3 secondary (2 points each), and 3 tertiary (1 point each)
see, at Le Bonheur Children’s Hospital or the Elvis Presley predictive criteria (Table 1). A maximum of 18 points is
Memorial Trauma Center between 1996 and 2013 were possible.
initially screened for study inclusion. Patients were eligible
if they suffered a firearm-induced, penetrating, intradural Statistical Analysis
traumatic brain injury and were 18 years old or younger. Bivariate analyses using chi-square and Fisher’s exact
Patients with multiple traumas, including concomitant ex- tests were performed to evaluate individual effects of each
tracranial penetrating injuries, were not included. Glasgow clinical and radiological criterion against overall survival:
Outcome Scale (GOS) score at the last available follow-up 1) ICP > 30 cm H2O; 2) bilateral fixed pupils; 3) systolic
was the primary outcome, with particular attention paid to blood pressure (SBP) < 100 mm Hg; 4) hematocrit < 30%;
a GOS score of 1, which is defined as death. 5) base deficit < -5 mEq/L; 6) international normalized
ratio (INR) > 1.5; 7) deep nuclear and/or third ventricular
Medical and Surgical Management injury; 8) mixed supra-/infratentorial injury; 9) injury to ≥
For patients in whom an intracranial pressure (ICP) 3 lobes, transventricular projectile trajectory; 10) bihemi-

596 J Neurosurg Pediatr Volume 17 • May 2016

Unauthenticated | Downloaded 04/21/24 01:20 PM UTC

 Management of pediatric intracranial gunshot wounds

TABLE 1. St. Louis Scale for Pediatric Gunshot Wounds to the

Head*
Predictive Criteria Description
Primary (3 points each) Bilateral fixed pupils on arrival; involve­
ment of deep nuclei &/or 3rd ven­tri­cle;
ICP >30 mm Hg
Secondary (2 points each) Mixed supra-/infratentorial involvement;
at least 3 lobes involved (single lobe
for cerebellum); transventricular injury
Tertiary (1 point each) Bihemispheric injury; SBP <100 mm Hg
on arrival; midline shift
SBP = systolic blood pressure.
* The points for each applicable criterion are summed for a total possible
score of 18. Modified from Bandt et al: J Neurosurg Pediatr 10:511–517, 2012.
Published with permission.

Fig. 1. Frequency of admission GCS scores in the Memphis cohort. ED

spheric injury; and 11) midline shift. For age comparison, = emergency department.
the cohort was divided into 2 groups: those 0–9 years and
those 10–18 years. An a value of 0.05 was chosen for statis-
tical significance. Statistical analysis was performed using ally fixed pupils and anisocoria were present in 52% (37 of
commercially available software (Prism 6.0, GraphPad). 71) and 14% (10 of 71) of patients, respectively.

Results Hospital Course

Demographics and Presentation Surgical intervention was performed in 39 patients
Seventy-one pediatric patients with an intracranial (54.9%). Eighteen patients underwent a craniotomy, in-
GSW presented to 1 of 2 Level 1 trauma centers between cluding debridement of skin, skull, dura, and brain. Re-
January 1996 and December 2013. The demographics of tained bone and metallic fragments that were easily acces-
this population are summarized in Table 2. The majority sible were removed. One patient developed a wound infec-
(57 [80.3%] of 71) were male, and the mean age at presen- tion requiring debridement. Twenty-one patients under-
tation was 14 years (range 19 months–18 years). The mean went decompressive craniectomy either preemptively (18
GCS score was 7.4 ± 5.0, with the frequency of scores on patients) or secondarily (3 patients) because of refractory
admission having a bimodal distribution (Fig. 1). Bilater- intracranial hypertension. Postoperative complications in
this group included 5 wound infections requiring reopera-
tion and 3 cerebrospinal fluid leaks.
TABLE 2. Patient demographics Among survivors (37 patients), the mean duration of
Characteristic Value (%) mechanical ventilation, stay in the intensive care unit, and
overall hospital stay was 11.3 days (range 0–44 days), 12.0
Sex days (range 1–33 days), and 14.8 days (range 2–44 days),
  Male 57 (80.3) respectively. Overall, 22 patients were discharged home
  Female 14 (19.7) and 15 patients were discharged to a rehabilitation facility
Race or unit.
   African American 34 (47.9)
Outcome
  White 27 (38)
The overall mortality rate in this series of patients was
  Hispanic 10 (14.1) 47.9% (34 of 71). Figure 2 shows the distribution of GOS
Age in yrs scores overall and among those who underwent surgical
  0–2 2 (3) intervention (39 patients), 2 of whom died postoperatively
  3–8 7 (10) (5.1%) and 30 (76.9%) of whom had a good clinical out-
  9–12 9 (13) come (GOS Score 4 or 5). As shown in Fig. 3, there was a
linear relationship between GCS score at presentation and
  13–15 21 (30)
GOS score at the last follow-up.
  16–18 32 (45) The clinical and radiological findings of the survivors,
Etiology stratified by either a poor (GOS Score 2 or 3, 7 patients) or
   Self-inflicted suicide 10 (14.1) a good (GOS Score 4 or 5, 30 patients) outcome, are listed
   Self-inflicted accidental 16 (22.5) in Table 3. In general, the children with a good clinical
   Assault by other 25 (35.2) outcome had fewer of the noted findings than those with
a poor outcome. It is difficult to draw meaningful conclu-
   Accidental by other 7 (9.8)
sions regarding outcome prediction, however, given the
  Bystander 13 (18.3) relative scarcity of these findings among survivors.

J Neurosurg Pediatr Volume 17 • May 2016 597

Unauthenticated | Downloaded 04/21/24 01:20 PM UTC

M. DeCuypere et al.

TABLE 3. Clinical and radiological criteria of survivors

GOS Score GOS Score
Variable 2 or 3 4 or 5
ICP >30 cm H2O 1 1
GCS score
  3–5 4 2
  6–8 3 4
Bilat fixed pupils 2 0
SBP <100 mm Hg 2 0
Hematocrit <30% 5 2
Base deficit <−5.0 mEq/L 2 0
INR >1.5 3 1
Deep nuclear/3rd ventricular injury 7 3
Mixed supra-/infratentorial injury 1 1
Fig. 2. Overall GOS scores of the study cohort and GOS scores after Injury to ≥3 lobes 7 4
surgical intervention. GOS scores: 1 = death; 2 = persistent vegetative Transventricular trajectory 5 0
state; 3 = dependent, severe disability; 4 = independent, minor disability; Bihemispheric injury 5 0
5 = normal, minimal disability.
Midline shift 7 12
Age in yrs
Predictors of Mortality
  0–9 1 4
Bivariate results are presented in Table 4. Nine of the
16 variables tested were predictive of death (p < 0.05).   10–18 6 26

St. Louis Scale for Pediatric Gunshot Wounds to the Head

value for death was 78.05%. The negative predictive value
The mean score on the St. Louis Scale for Pediatric (St. Louis score ≤ 4 indicative of survival) was 93.33%.
Gunshot Wounds to the Head in children who survived
and those who died was 2.8 (range 0–12) and 10 (range
3–14), respectively. Table 5 shows the average scores and Discussion
range for the survivors by GOS score. Interestingly, there Our Results
were 9 survivors (24.3%) in the cohort whose St. Louis Children are not immune to gunshot injuries, inten-
score was ≥ 5. All of these patients underwent surgery at tional or accidental, that plague many cities across the
the time of presentation. Three of these patients had a GOS US. When presented with a child who has sustained an
score of 4 (favorable outcome), 5 patients had a GOS score intracranial GSW, the neurosurgeon must quickly decide
of 3, and 1 patient had a GOS score of 2 at last follow-up. whether the child has a fatal injury, an injury that is poten-
The results of our evaluation of the St. Louis scale as a tially nonfatal but very likely to have a devastating neu-
diagnostic test are detailed in Table 6. For the purposes of rological outcome, or a survivable injury for which there
evaluation, a score ≥ 5 was considered a positive test pre- is a reasonable chance of maintaining or regaining mean-
dictive for death. The sensitivity of the scale was 94.12%, ingful neurological function. The physician’s appraisal of
while the specificity was 75.68%. The positive predictive the neurological exam and imaging findings are then pre-
sented to the family and, in alignment with their wishes, a
decision can be made on how to proceed with treatment.
This process of determining what to do is more challeng-
ing in children than in adults owing to the former’s young
age and greater propensity to recover neurological func-
tion. This latter point is highlighted by the differences in
brain death criteria between children and adults.21
Our series is one of the few that focuses on GSWs in
children. Many reports of civilian cranial GSWs contain
both pediatric and adult patients. We refer the reader to
the paper by Bandt et al.,2 which includes a tabulated sum-
mary of key articles. In their series of 82 children, Beaver
et al.3 concluded that a child who has sustained a firearm
injury is more likely to know the perpetrator, to be killed
in the home by a readily available unsecured firearm, and
to die of severe head injury. Paret et al. reported on 51
children and found admission GCS scores, CT findings
of intraventricular hemorrhage and midline shift, and
Fig. 3. Relationship of admission GCS score and overall GOS score. metabolic abnormalities to be of prognostic value.22 One

598 J Neurosurg Pediatr Volume 17 • May 2016

Unauthenticated | Downloaded 04/21/24 01:20 PM UTC

 Management of pediatric intracranial gunshot wounds

TABLE 4. Bivariate testing of clinical and radiological criteria TABLE 6. Evaluation of the St. Louis scale
Variable Survived Died NNT p Value Statistic Value 95% CI
ICP >30 cm H2O 2 4 16 0.4077 Sensitivity 94.12% 80.29%–99.11%
GCS score* Specificity 75.68% 58.80%–88.20%
  3–5 6 31 2 <0.0001† Positive likelihood ratio* 3.87 2.18–6.87
  6–8 7 3 5 0.2219 Negative likelihood ratio† 0.08 0.02–0.30
Bilat fixed pupils 2 25 2 <0.0001† Death prevalence 47.89% 35.88%–60.08%
SBP <100 mm Hg 2 13 4 0.009† PPV 78.05% 62.38%–89.42%
Hematocrit <30% 4 13 4 0.006† NPV 93.33% 77.89%–98.99%
Base deficit <−5.0 mEq/L 5 15 4 0.0038† NPV = negative predictive value; PPV = positive predictive value.
INR >1.5 2 3 12 0.0962 * Positive test defined as St. Louis score ≥ 5.
Deep nuclear/3rd ventricular 10 26 2 <0.0001† † Negative test defined as St. Louis score ≤ 4.
injury
Mixed supra-/infratentorial injury 2 6 8 0.1346 nificant association between a GCS score ≤ 5 and over-
Injury to ≥3 lobes 10 19 4 0.0084† all mortality, we do not recommend using a strict GCS
Transventricular trajectory 5 26 2 <0.0001†
cutoff score alone because of the potential inaccuracy of
the admission GCS score in a patient whose neurological
Bihemispheric injury 5 26 2 <0.0001† function is depressed iatrogenically or in response to inad-
Midline shift 19 22 7 0.1697 equate volume resuscitation or hypothermia. In agreement
Age in yrs with other authors, we found a strong association between
  0–9 5 5 41 0.8853 bilaterally fixed pupils and death.10,23,26 The significant as-
  10–18 32 29 41 0.8853 sociation of anemia (hematocrit < 30%) and intravascular
volume depletion (base deficit < -5 mEq/L) with mortal-
NNT = number needed to treat. ity reflects the critical importance of rapid correction with
* All patients with GCS score > 8 survived. fluid and/or blood. Coagulopathy at presentation (INR >
† Statistically significant.
1.5) was not significantly associated with death, probably
given the fact that temporary coagulopathy is not as po-
of the largest series known to us is by Levy et al., which tentially damaging to the at-risk brain as hypotension and
comprised 105 children treated at the University of South- that the coagulopathy can usually be readily reversed with
ern California/Los Angeles County Medical Center.17 The transfusion of plasma. The lack of an association between
majority of the children in that series (72%) were treated intracranial hypertension (ICP > 30 cm H2O) and death
for gang-related injuries, and patient age, sex, GCS score, is due to inadequate data given our bias for upfront sur-
projectile entry site, and bihemispheric injury were pre- gery versus ICP monitoring. Although preoperative ICP
dictive variables. monitoring was rarely performed, it was routinely used
We found that certain clinical (admission GCS score postoperatively.
≤ 5, bilateral fixed and dilated pupils), laboratory (initial Similar to the radiological criteria set forth in the St.
hematocrit < 30%, base deficit < -5 mEq/L), and CT (deep Louis scale, our data support the ominous findings of
nuclear/third ventricular injury, injury to ≥ 3 lobes, trans- deep nuclear/third ventricular injury and transventricu-
ventricular injury, bihemispheric injury) results were sta- lar trajectory. The predictive value of transventricular
tistically associated with death. An age less than 9 years, and transhemispheric missile injuries has been reported
an initial ICP > 30 cm H2O (or mm Hg), both supra- and extensively in the literature on adults.4,5,8,13,14,25,26,28 Mixed
infratentorial injury, and midline shift were not predictive. supratentorial/infratentorial injury did not reach statisti-
Several other reports emphasize the importance of the cal significance—again because of a lack of data—but it
GCS score at presentation in management decision mak- should be noted that this injury was associated with 75%
ing.5,7,8,10,16,17,22,23,26 While our data do demonstrate a sig- mortality. In contrast to Bandt et al.,2 we found that mid-
line shift was not associated with mortality. This finding
may be attributable to the fact that this criterion was a bi-
TABLE 5. St. Louis Scale for Pediatric Gunshot Wounds to the nary variable (yes/no) as opposed to a cutoff value. For ex-
Head scores in 71 patients with GSWs ample, a midline shift > 5 mm is a more significant finding
Group Mean Score (range) than a shift of 1–2 mm. Furthermore, as stated previously,
our threshold in performing a craniectomy was low, which
Survivors 2.8 (0–12) meant that children with midline shift were often rapidly
   GOS Score 2 12 (NA) taken for decompression, which would have relieved the
   GOS Score 3 8.5 (0–12) midline shift and intracranial hypertension to a greater
   GOS Score 4 2.8 (0–9) degree than medical management.
   GOS Score 5 0.4 (0–2)
Another important but often immeasurable factor in
determining the extent of tissue damage and neurological
Nonsurvivors 10 (3–14) injury is the muzzle energy, which is the kinetic energy of
NA = not applicable. the projectile upon exiting the barrel of the weapon. It is

J Neurosurg Pediatr Volume 17 • May 2016 599

Unauthenticated | Downloaded 04/21/24 01:20 PM UTC

M. DeCuypere et al.

directly proportional to the weight (or “grain”) of the bul- nately, the more difficult decisions involve the children
let, but more importantly the velocity as kinetic energy = whose St. Louis scale score is greater than 5.
1/2 × mass × velocity2. The greater energy the projectile
has upon impact, the greater energy that is dissipated to Study Limitations
the surrounding tissue, causing cavitation injury. It is rare A retrospective study design is always touted as a limi-
that the actual bullet is recovered from the patient, and if tation, but it is difficult to conduct a prospective single-
it is, it is often a small-caliber bullet (0.22 to 0.38 caliber) center study on an event that occurs with an average fre-
from a low-muzzle-velocity weapon (900 to 1300 ft/sec).5 quency of 4 patients per year. Even though our series is
In the pediatric population, published mortality rates one of the larger ones, it still has a relatively small number
from intracranial GSWs range from 20% to 60%.2,7,11,16,17,19,22 of patients, which limits the strength of our findings and
The mortality rate of our cohort fell within this range conclusions.
(47.9%) and was lower than the mortality reported by Bandt
and colleagues (65%).2 Unfortunately, those authors did not
provide a breakdown of the proportion of patients under- Conclusions
going surgery. In our cohort, 54.9% of patients underwent In Memphis and other large metropolitan communities,
surgical intervention. Of the survivors, 81% showed only pediatric intracranial GSWs are an unfortunate reality.
minor disability or better at the last follow-up (GOS Score In this study, 52% of children with an intracranial GSW
4 or 5). This rate is comparable with the results from the St. survived with aggressive surgical and medical care, and
Louis study in which 88% of survivors had a GOS score 81% of survivors had a favorable outcome (GOS Score 4
of 4 or 5.2 or 5) at the last follow-up. The utility of several clinical,
laboratory, and radiological prognostic factors were veri-
St. Louis Scale for Pediatric Gunshot Wounds to the Head fied. None of these factors, with the exception of bilateral
A number of prognostic scales have been proposed to fixed and dilated pupils, should be used alone to determine
help in determining the viability of patients with cranial prognosis and treatment, but instead all factors should be
GSWs.8,9,18,28,30 The St. Louis scale is the only one that considered together as they have greater predictive power
we know of that is specific for children and marks a sig- collectively. When used in our population as a predictive
nificant contribution to the literature.2 In Bandt and col- tool, the St. Louis Scale for Pediatric Gunshot Wounds to
leagues’ series, a St. Louis scale score ≤ 4 was associated the Head was superior at forecasting survival after a gun-
with a positive predictive value of 88.9% for survival; a shot injury rather than death.
score ≥ 5 was associated with a negative predictive value
of 96.7% for death. The sensitivity of the scale, using the Acknowledgments
St. Louis cohort, was 93.55% and the specificity of the We thank Andrew J. Gienapp for technical and copyediting
scale was 94.12%. preparation of the manuscript and figures and for publication assis-
The mean St. Louis score for the nonsurvivors in our tance with this manuscript.
cohort was 10, which is in general agreement with the
utility of the scale. However, the mean St. Louis score for
the group with a GOS score of 3 was 8.5, which demon- References
strates that some patients with a St. Louis score ≥ 5 can 1. Aarabi B, Tofighi B, Kufera JA, Hadley J, Ahn ES, Cooper C,
and will survive. In fact, 9 survivors had St. Louis scores et al: Predictors of outcome in civilian gunshot wounds to the
head. J Neurosurg 120:1138–1146, 2014
≥ 5. Of these 9 patients, 3 had a GOS score of 4 at the 2. Bandt SK, Greenberg JK, Yarbrough CK, Schechtman KB,
last follow-up and 5 had a GOS score of 3. While severe Limbrick DD, Leonard JR: Management of pediatric intra-
disability may be considered an unfavorable outcome by cranial gunshot wounds: predictors of favorable clinical out-
some (and probably more so for an adult patient), many come and a new proposed treatment paradigm. J Neurosurg
families would disagree with this sentiment when faced Pediatr 10:511–517, 2012
with the imminent death of their child. Upon further in- 3. Beaver BL, Moore VL, Peclet M, Haller JA Jr, Smialek J,
vestigation into this subgroup of the cohort, the higher St. Hill JL: Characteristics of pediatric firearm fatalities. J Pedi-
Louis scores were primarily attributable to deep nuclear atr Surg 25:97–100, 1990
injury and bihemispheric injury without transventricular 4. Cavaliere R, Cavenago L, Siccardi D, Viale GL: Gunshot
wounds of the brain in civilians. Acta Neurochir (Wien)
trajectory. Additional points were gained from low SBP 94:133–136, 1988
and midline shift. As such, some survivors with the poten- 5. Clark WC, Muhlbauer MS, Watridge CB, Ray MW: Analysis
tial for a favorable outcome may be missed if judged by of 76 civilian craniocerebral gunshot wounds. J Neurosurg
the St. Louis scale score alone. 65:9–14, 1986
Using our patient population, we determined that the St. 6. Coronado VG, Xu L, Basavaraju SV, McGuire LC, Wald
Louis scale was sensitive (94.12%), but lacked specificity MM, Faul MD, et al: Surveillance for traumatic brain injury-
for death as an outcome (75.68%). A score ≥ 5 was associ- related deaths—United States, 1997-2007. MMWR Surveill
ated with a positive predictive value of 78.05% for death; Summ 60:1–32, 2011
7. Coughlan MD, Fieggen AG, Semple PL, Peter JC: Cranio-
a score ≤ 4 was associated with a negative predictive value cerebral gunshot injuries in children. Childs Nerv Syst
of 93.33% for death (in other words 93.33% predictive for 19:348–352, 2003
survival). Therefore, we conclude that the St. Louis scale is 8. Grahm TW, Williams FC Jr, Harrington T, Spetzler RF: Ci-
more useful for determining those who will survive (score vilian gunshot wounds to the head: a prospective study. Neu-
≤ 4) rather than those who will die (score ≥ 5). Unfortu- rosurgery 27:696–700, 1990

600 J Neurosurg Pediatr Volume 17 • May 2016

Unauthenticated | Downloaded 04/21/24 01:20 PM UTC

 Management of pediatric intracranial gunshot wounds

9. Gressot LV, Chamoun RB, Patel AJ, Valadka AB, Suki D, 23. Petridis AK, Doukas A, Barth H, Mehdorn M: Outcome of
Robertson CS, et al: Predictors of outcome in civilians with craniocerebral gunshot injuries in the civilian population.
gunshot wounds to the head upon presentation. J Neurosurg Prognostic factors and treatment options. Cent Eur Neuro-
121:645–652, 2014 surg 72:5–14, 2011
10. Hofbauer M, Kdolsky R, Figl M, Grünauer J, Aldrian S, Os- 24. Prasad MR, Ewing-Cobbs L, Swank PR, Kramer L: Predic-
termann RC, et al: Predictive factors influencing the outcome tors of outcome following traumatic brain injury in young
after gunshot injuries to the head—a retrospective cohort children. Pediatr Neurosurg 36:64–74, 2002
study. J Trauma 69:770–775, 2010 25. Selden BS, Goodman JM, Cordell W, Rodman GH Jr,
11. Irfan FB, Hassan RU, Kumar R, Bhutta ZA, Bari E: Cranio- Schnitzer PG: Outcome of self-inflicted gunshot wounds of
cerebral gunshot injuries in preschoolers. Childs Nerv Syst the brain. Ann Emerg Med 17:247–253, 1988
26:61–66, 2010 26. Shaffrey ME, Polin RS, Phillips CD, Germanson T, Shaffrey
12. Joseph B, Aziz H, Pandit V, Kulvatunyou N, O’Keeffe T, CI, Jane JA: Classification of civilian craniocerebral gunshot
Wynne J, et al: Improving survival rates after civilian gun- wounds: a multivariate analysis predictive of mortality. J
shot wounds to the brain. J Am Coll Surg 218:58–65, 2014 Neurotrauma 9 (Suppl 1):S279–S285, 1992
13. Kaufman HH: Civilian gunshot wounds to the head. Neuro- 27. Siccardi D, Cavaliere R, Pau A, Lubinu F, Turtas S, Viale
surgery 32:962–964, 1993 GL: Penetrating craniocerebral missile injuries in civilians: a
14. Kim KA, Wang MY, McNatt SA, Pinsky G, Liu CY, Gian- retrospective analysis of 314 cases. Surg Neurol 35:455–460,
notta SL, et al: Vector analysis correlating bullet trajectory to 1991
outcome after civilian through-and-through gunshot wound 28. Stone JL, Lichtor T, Fitzgerald LF: Gunshot wounds to the
to the head: using imaging cues to predict fatal outcome. head in civilian practice. Neurosurgery 37:1104–1112, 1995
Neurosurgery 57:737–747, 2005 29. Pruitt BA Jr (ed): Surgical management of penetrating brain
15. Krieger MD, Levy ML, Apuzzo ML: Gunshot wounds to the injury. J Trauma 51 (2 Suppl):S16–S25, 2001
head in an urban setting. Neurosurg Clin N Am 6:605–610, 30. Turina D, Sustić A, Tićac Z, Dirlić A, Krstulović B, Glavas
1995 A, et al: War head injury score: an outcome prediction model
16. Levy ML: Outcome prediction following penetrating cranio- in War casualties with acute penetrating head injury. Mil
cerebral injury in a civilian population: aggressive surgical Med 166:331–334, 2001
management in patients with admission Glasgow Coma Scale 31. Zafonte RD, Wood DL, Harrison-Felix CL, Valena NV,
scores of 6 to 15. Neurosurg Focus 8(1):e2, 2000 Black K: Penetrating head injury: a prospective study of out-
17. Levy ML, Masri LS, Levy KM, Johnson FL, Martin-Thom- comes. Neurol Res 23:219–226, 2001
son E, Couldwell WT, et al: Penetrating craniocerebral injury
resultant from gunshot wounds: gang-related injury in chil-
dren and adolescents. Neurosurgery 33:1018–1025, 1993
18. Martins RS, Siqueira MG, Santos MT, Zanon-Collange N, Disclosures
Moraes OJ: Prognostic factors and treatment of penetrating
The authors report no conflict of interest concerning the materi-
gunshot wounds to the head. Surg Neurol 60:98–104, 2003
als or methods used in this study or the findings specified in this
19. Miner ME, Ewing-Cobbs L, Kopaniky DR, Cabrera J,
paper.
Kaufmann P: The results of treatment of gunshot wounds to
the brain in children. Neurosurgery 26:20–25, 1990
20. Murano T, Mohr AM, Lavery RF, Lynch C, Homnick AT, Author Contributions
Livingston DH: Civilian craniocerebral gunshot wounds: Conception and design: all authors. Acquisition of data: DeCuy-
an update in predicting outcomes. Am Surg 71:1009–1014, pere. Analysis and interpretation of data: Klimo, DeCuypere.
2005 Drafting the article: Klimo, DeCuypere. Critically revising the
21. Nakagawa TA, Ashwal S, Mathur M, Mysore M: Clinical article: all authors. Reviewed submitted version of manuscript: all
report—Guidelines for the determination of brain death in authors. Approved the final version of the manuscript on behalf
infants and children: an update of the 1987 task force recom- of all authors: Klimo. Study supervision: Klimo.
mendations. Pediatrics 128:e720–e740, 2011
22. Paret G, Barzilai A, Lahat E, Feldman Z, Ohad G, Vardi Correspondence
A, et al: Gunshot wounds in brains of children: prognostic Paul Klimo Jr., Semmes-Murphey Neurologic & Spine Institute,
variables in mortality, course, and outcome. J Neurotrauma 6325 Humphreys Blvd., Memphis, TN 38120. email: pklimo@
15:967–972, 1998 semmes-murphey.com.

J Neurosurg Pediatr Volume 17 • May 2016 601

Unauthenticated | Downloaded 04/21/24 01:20 PM UTC