Injury, Int. J.

Care Injured 46 (2015) 49–53

Contents lists available at ScienceDirect

Injury
journal homepage: www.elsevier.com/locate/injury

The selective conservative management of penetrating thoracic

trauma is still appropriate in the current era
Victor Y. Kong a,*, Benn Sartorius b,1, Damian L. Clarke a,2
a
Pietermaritzburg Metropolitan Trauma Service, Department of Surgery, University of KwaZulu Natal, Pietermaritzburg 3216, South Africa
b
Discipline of Public Health Medicine, School of Nursing and Public Health, University of KwaZulu-Natal, Durban, South Africa

A R T I C L E I N F O A B S T R A C T

Article history: Introduction: Traumatic pleural collections secondary to penetrating chest trauma are generally
Accepted 11 July 2014 managed by intercostal chest drainage (ICD), but these protocols were developed a few decades ago
when stabs (SWs) predominated over gunshot wounds (GSWs). This study reviews the outcome of a
Keywords: selective conservative approach to penetrating thoracic trauma to establish if it is still appropriate in the
Penetrating current era.
Thoracic trauma Materials and methods: We reviewed 827 patients over a four-year period with penetrating unilateral
Stab
non-cardiac wounds of the chest in order to review the efficacy of our policy and to define the differences
Gunshot
in the spectrum of injury between SWs and GSWs.
Results: Ninety-two per cent (764/827) were males, and the median age was 24 years. Seventy-six per
cent (625/827) sustained SWs and twenty-four per cent (202/827) GSWs. Chest pathologies were:
pneumothorax (PTX): 362 (44%), haemothorax (HTX): 150 (18%) and haemopneumothorax (HPTX): 315
(38%). Ninety-six per cent of patients were managed non-operatively. Four per cent (36/827) were
subjected to a thoracotomy [31 SWs and 5 GSWs]. No difference was observed in terms of the need for
operative intervention: 5% vs. 3% [p = 0.202]. PTX was seen exclusively in SWs: 58% vs. 0% and there were
significantly more HPTXs seen in the GSWs: HPTX: 24% vs. 81% [p < 0.001]. The median days of ICD in
situ were significantly longer in GSWs compared to SWs for all pathologies. For HTX: 4.5 (interquartile
range [IQR]: 3–6) vs. 3.5 (IQR: 0–5) days, p = 0.001 and HPTX: 4 (IQR: 3–5) vs. 3.0 (IQR: 3–4) days,
p < 0.001. There were seven (15%) complications. A total of five (13%) patients died and all deaths were
confined to the operative group.
Conclusions: SWs continue to predominate over GSWs. PTXs were more commonly associated with SWs,
whilst HPTX are more commonly associated with GSWs. A policy of selective conservatism is still
applicable to the management of traumatic pleural collections.
ß 2014 Elsevier Ltd. All rights reserved.

Introduction entrenched in South African trauma centres, it is apparent that

Traumatic pleural collections secondary to penetrating chest
trauma are frequently encountered in trauma centres around the
world [1–3] and remain amongst the most common injuries
managed at our institution [4]. Previous local studies from over
two decades ago reported a high incidence of thoracic stab injuries
(SWs) [5,6], whilst more recent studies from the same institution
have shown an increase in thoracic gunshot injuries (GSWs) [3,7].
Although the philosophy of selective conservatism is firmly
GSWs of the chest are more lethal than SWs and are associated with a
different spectrum of injuries [7]. This study confines itself to
traumatic pleural collections secondary to unilateral penetrating
chest trauma in order to better define the differences in the spectrum
of injury between the two mechanisms and to review whether
management protocols derived from a historical experience heavily
weighted towards the management of SWs are still appropriate.
Materials and methods

Setting
* Corresponding author. Tel.: +27 33 395 4911.
E-mail addresses: victorywkong@yahoo.com (V.Y. Kong), sartorius@ukzn.ac.za
(B. Sartorius), damianclar@gmail.com (D.L. Clarke).
This was a retrospective study undertaken in the Pietermaritz-
1
Tel.: +27 31 260 4459. burg Metropolitan Trauma Service (PMTS), Pietermaritzburg,
2
Tel.: +27 33 395 4911. South Africa. A retrospective review of a prospectively maintained

http://dx.doi.org/10.1016/j.injury.2014.07.011
0020–1383/ß 2014 Elsevier Ltd. All rights reserved.
50 V.Y. Kong et al. / Injury, Int. J. Care Injured 46 (2015) 49–53

regional trauma registry was performed over a four-year period

from January 2010 to December 2013. Ethics approval for this
study and for the maintenance of the registry was formally
endorsed by the Biomedical Research Ethics Committee (BREC) of
the University of KwaZulu Natal (reference number: 207/09). The
PMTS provides definitive trauma care to the city of Pietermaritz-
burg and the surrounding rural areas with a total catchment
population of over three million people. It also serves as the
definitive trauma referral service for nineteen other district
hospitals within the province [8]. All patients who sustained
isolated penetrating injuries to the chest were eligible for
inclusion. Patients with trans-mediastinal injuries, penetrating
cardiac trauma and concurrent injuries above and below the
diaphragm were excluded from this study.

Mortuary data

All patients who suffer an unnatural death in South Africa are

mandated by law to undergo a forensic autopsy. The administra-
tion of the medico-legal laboratories is under the jurisdiction of the
South African Police Services and as such, data are not readily
available to medical staff. We have instituted a number of
strategies to access crude data from the medico-legal services
and have since 2006 received aggregated annual reports on
the load at the police mortuary in Pietermaritzburg. Refining the
reporting of these data and agreeing on appropriate data sets is
the subject of on-going negotiation between the PMTS and the
medico-legal services.

Management protocols Fig. 1. HTX following stab injuries requiring an ICD.

The management principles of thoracic trauma at our

institution are based on a policy of selective conservatism. All
patients with penetrating thoracic trauma who present to the
service, are assessed by the duty trauma residents. The residents
are overseen by a group of specialist trauma surgeons. Following
clinical assessment, a chest radiograph (CXR) is obtained to
confirm the presence of a traumatic pleural collection. All
patients with significant pleural collections, which fulfil our
criteria immediately receive an ICD. Our criteria for ICD insertion
include a pneumothorax with a visible rim >2 cm measured from
the apex to the cupula, and/or a fluid level above the seventh rib.
All ICDs are inserted in the trauma unit under local anaesthesia
according to the Advanced Trauma Life Support (ATLS) guidelines
and the patients are then admitted to the surgical ward. Fig. 1
shows an example of a typical young male with a significant HTX
following a stab to the right chest, which required an ICD. Our
ICD removal criteria are as follows: (1) no evidence of clinical
distress, (2) no swing of fluid level in drain bottle, and (3)
drainage <50 ml per day. Patients are repeatedly encouraged to
mobilise vigorously, by climbing the stairs of our seven floor
main hospital building. Daily visits by physiotherapists are
routine.
Indications for operative exploration
We use the following criteria in conjunction with the patient’s
overall clinical picture as indications for emergency surgical
exploration: (1) on-going haemodynamic instability, (2) massive
haemothorax with initial ICD output >1500 ml blood, not
responding to resuscitation, (3) on-going ICD drainage of blood
>200 ml per hour, with no signs of clinical improvement or
worsening clinical status. (4) Clinical evidence of cardiac
tamponade in shocked patients and (5) imminent cardiac
arrest from any penetrating mechanism. We do have facilities
for auto-transfusion and we make use of this facility on a regular
basis in patients who undergo operative exploration [9].
Data collection
Data collected included: basic demographics, mechanism of
injury and side of injury. Further details on the initial pleural
collections identified were recorded. We divided these into
pneumothorax (PTX), haemothorax (HTX), or haemopneu-
mothorax (HPTX). The initial management and the proportion of
patients managed exclusively with ICDs were recorded. All
patients who required emergency thoracic exploration were
reviewed and further analysed. All data were extracted from the
registry onto an EXCELß spread sheet for processing.
Data analysis
Data were analysed using Stata 13.0 SE (StataCorp, Stata
Statistical Software: Release 13, College Station, TX: StataCorp LP,
2013). Continuous variables were summarised using the mean and
standard deviation. If data were skewed then medians and
interquartile ranges were presented. Categorical data were
represented using frequency tables. The non-parametric equiva-
lent of the t-test, namely the Wilcoxon rank-sum test (also known
as the Mann–Whitney two-sample statistic), was used to identify
significant rank (median) differences in continuous explanatory
variables by participation status as the data were not normally
distributed. Categorical explanatory variables were cross tabulated
against by injury type (GSWs vs. SWs) and significant association
was identified using the standard Pearson’s chi-square (x2) test. If
the expected cell counts in the cross tabulation were less than 5
then the Fisher’s exact test was preferred. A p-value of less than
0.05 was deemed statistically significant.
 V.Y. Kong et al. / Injury, Int. J. Care Injured 46 (2015) 49–53
Results
Demographics
Over the four-year study period, a total of 827 patients with
unilateral penetrating thoracic trauma were identified. Ninety-two
per cent (764/827) were males, with an overall median age of 24
years (interquartile range [IQR]: 20–29).
Mechanisms
Seventy-six per cent (625/827) sustained SWs and the
remaining twenty-four per cent (202/827) sustained GSWs.
Sixty-four per cent (530/827) of all injuries were sustained on
the left side, and thirty-six per cent (297/827) were on the right.
Initial pathologies were as follows: PTXs: 362 (44%), HTXs: 150
(18%) and HPTXs: 315 (38%). Four per cent (36/827) of all patients
had indications for operation and were subjected to emergency
exploration. The remaining 96% were successfully managed non-
operatively. Fig. 2 summarises the relative proportion by mecha-
nism of injury.
Stab injuries
Of the 625 patients who sustained SWs, 91% (568/625) were
males, and the overall median age of all patients was 24 (IQR: 20–
29) years. Sixty-four per cent (397/625) of injuries were left and
thirty-six per cent (228/625) were on the right. There were: 362
PTXs (58%), 113 HTXs (18%) and 150 (24%) HPTXs. Thirty-one of
these 625 patients (5%) required operations. Of the 594 who did
not have operations, the median lengths of ICD in situ (in days)
were: PTX: 1 (IQR: 1–1) day, HTX: 4 (IQR: 3–5) days, and HPTX: 4
(IQR: 3–4) days.
GSWs
Of the 202 patients who sustained GSWs, 97% (196/202) were
males and the median age for all patients was 25 (IQR: 20–28)
years. Sixty-six per cent (133/202) of injuries were left sided and
the remaining thirty-four per cent (69/202) were right sided. There
were no isolated PTXs, 37 (18%) HTXs, and 165 (82%) HPTXs. Five of
Fig. 2. Mechanism of injury in the 827 patients.
51
the 202 patients (2%) required operations. Of the 197 who did not
have operations, the median lengths of ICD in situ were: HTX: 5
(IQR: 4–6) days, and HPTX: 4 (IQR: 3–5) days.
SWs vs. GSWs
There were significantly more males amongst the GSW group
(91% vs. 97%) [p = 0.007]. There was no difference in the side of
injury between SWs and GSWs: left side: 63% vs. 66%, right side:
37% vs 34% [p = 0.608]. No difference was observed in terms of the
need for operative intervention: 5% vs 3% [p = 0.191]. Table 1 shows
a comparison of basic demographics and patterns of injury
between SWs vs. GSWs injuries.
In terms of thoracic pathologies, PTX was seen exclusively in
SWs: 58% vs. 0%. There were significant more HPTXs seen in the
GSWs group: HPTX: 24% vs. 81% [p < 0.001]. There is no significant
difference in HTXs between SWs and GSWs. Table 2 compares the
proportion of different types of thoracic pathology in both groups.
The median days of ICD in situ were significantly longer in
GSWs for all pathologies. For HTX: 3.5 (IQR: 0–5) vs. 4.5 (IQR: 3–6)
days, p = 0.001. For HPTX: 3 (IQR: 3–4) vs. 4 (IQR: 3–5) days,
p < 0.001. Table 3 compares the median ICD days between SW and
GSW patients.
Operations
A total of 36 (4%) patients underwent operative exploration.
Ninety-seven per cent (35/36) of all patients who underwent
operative exploration were males and the median age for all
patients was 23 (IQR: 20–29) years. Seventy-two per cent (26/36)
were left sided injuries and twenty-eight per cent (10/36) were
right sided. Surgical approaches used were: 20 (56%) median
sternotomy, 14 (39%) anterolateral thoracotomy, and 2 (6%)
clamshell thoracotomy.
Ninety-seven per cent (35/36) of all patients who were
operated on had positive findings. One patient underwent a
negative thoracotomy. He presented severely inebriated with
multiple left sided SWs and with hypotension, which did not
respond to fluid resuscitation. He was in imminent cardiac arrest in
OR, when an anterolateral thoracotomy was performed, for what
was essentially an undiagnosed tension pneumothorax. He made
52 V.Y. Kong et al. / Injury, Int. J. Care Injured 46 (2015) 49–53

Table 1 Table 3
Basic demographics and pattern of injuries. Median ICD days by injury mechanism.

SWs % GSWs % p valuea Pathology Median ICD days (IQRa) p valueb

N = 625 N = 202 SWs GSWs
Median age in years 24 24 0.339 PTX 1 (1–1) 0 NA
(IQRb) (20–29) (20–28) HTX 3.5 (0–5) 4.5 (3–6) 0.001
HPTX 3 (3–4) 4 (3–5) <0.001
Gender
Male 568 91 196 97 0.007 PTX, pneumothorax; HTX, haemothorax; HPTX, haemopneumothorax.
Female 57 9 6 3 a
Interquartile range.
b
Wilcoxon rank-sum test.
Side
Left 397 64 133 66 0.608
Right 228 36 69 34

Operation 31 5 5 2 0.191 from 2006 to 2009. Table 5b shows the data from 2010 to 2011 for
No operation 594 95 197 98 all penetrating torso trauma.
a
Wilcoxon rank-sum test for median comparison by injury group, chi-square
(x2) test for comparison of proportions by injury group.
b
Interquartile range.
Discussion

Traumatic pleural collections secondary to penetrating chest

an uneventful recovery and was discharged alive. Of the 35
patients with positive findings at thoracotomy, 31 had a single
structure injured, whilst 4 had more than one structure injured.
There were a total of 39 intra-thoracic injuries identified in 35
patients. These were: major pulmonary laceration: 20, subclavian
vein: 10, internal thoracic artery injury: 4, pulmonary hila injury:
3, aortic arch: 1, and brachiocephalic artery: 1. Fig. 3a and b shows
a 30-year-old patients with a GSW casing laceration to the right
lung.
Of the four patients who sustained injuries to more than one
structure, the injuries were as follows: three had a combined
pulmonary laceration and hilar injuries, and one had a combined
pulmonary laceration and subclavian vein injury. Table 4 sum-
marises the 39 structures injured in the 35 patients.
trauma are common in South Africa [3–7]. Over three decades ago,
Muckart et al. at our parent institution King Edward VIII Hospital in
Durban, noted that over half of all intensive care unit admissions
were trauma related and that a total of 1662 patients with
penetrating thoracic trauma were managed in a single year at that
institution [3]. A similar study by Demetriades et al. from
Johannesburg in 1986 described 543 patients with thoracic SWs
managed over a fifteen-month period [10]. During the next two
decades, the overall incidence of penetrating thoracic injuries

Morbidity and mortality

Eighty-four per cent (32/36) of all thirty-six patients required

post-operative intensive care (ICU) admission. There were seven
(15%) complications in the operative group: five developed
ventilator associated pneumonia, one developed central line
related sepsis, and one had sternal sepsis. A total of five (13%)
patients died and all of these deaths were confined to the operative
group (two SWs and three GSWs). Four patients died in OR and one
was admitted to ICU but died later from multiple organ failure.

Mortuary data

Tables 5a and 5b summarise available mortuary data for the

period 2006–2011. As there is no simple mechanism to obtain the
data, the data set varies over the five-year period. GSWs are
associated with significantly higher mortality rates than SWs.
Table 5a shows the mortuary data for penetrating thoracic trauma

Table 2
Different types of thoracic pathology.

SWs % GSWs % p valuea

N = 625 N = 202

PTX 362 58 0 0 <0.001

HTX 113 18 37 18
HPTX 150 24 165 82

PTX, pneumothorax; HTX, haemothorax; HPTX, haemopneumothorax. Fig. 3. (a) GSW of the right lung. (b) Through-through injury of the right lung, with
a
Fisher’s exact test. profuse haemorrhage.
 V.Y. Kong et al. / Injury, Int. J. Care Injured 46 (2015) 49–53 53

Table 4 managed with an ICD [2,7,13]. Thoracic GSWs are more likely to
Intra-thoracic injuries in 35 patients.
cause a combination injury in the form of a HPTX [13]. Whilst these
Injured structure N = 39 % injuries are also treated conservatively, they require longer periods
Pulmonary laceration 20 51 of ICD drainage and are more prone to develop complications.
Subclavian vein 10 26
Internal thoracic artery 4 10
Pulmonary hilum 3 8 Conclusions
Aortic arch 1 3
Brachiocephalic artery 1 3 The management algorithms for traumatic pleural collections
continue to be based on closed intercostal chest drainage. Despite
the relative increase in GSW as a mechanism for penetrating
Table 5a
thoracic trauma our algorithms appear to be appropriate. A defined
Mortuary data for penetrating thoracic trauma from 2006 to 2009. small subset of patients will require operation and if we improve
our retrieval systems we may find the rate of operative exploration
Authora Mechanism Total Mortuary Mortality
for both thoracic GSWs and SWs increase. The spectrum of
[N = 1862] [N = 676] (%)
traumatic pleural collections differs according to the mechanism of
Clarke et al. [16] GSWs 259 135 52
injury. Simple PTXs are more commonly associated with SWs,
2006–2009 SWs 1603 541 34
whilst complex HPTX are more commonly associated with GSWs.
1862 676 36

SWs, stab wounds; GSWs, gunshot wounds.

a Conflict of interest
These data refer specifically to penetrating wounds of the chest.

There are no financial and personal relationships with other

people or organisations that could inappropriately influence (bias)
Table 5b their work.
Mortuary data for penetrating torso trauma from 2010–2011.

Mechanism Total Mortuary Mortality References

[N = 1307] [N = 267]
GSWs 223 95
2010–2011 SWs 1084 172
Unpublished data 1307 267
SWs, stab wounds; GSWs: gunshot wounds.
remained constant, but the mechanism of the injury changed [6,7].
Muckart et al. in 1995 noted that although there was a 30%
decrease in thoracic SWs, there was a dramatic increase in the
number of thoracic GSWs in the order of 800% [3]. At the turn of the
millennium, Madiba et al. from the same institution published his
review of 589 patients and found that 80% were SWs and 20%
GSWs [7]. Our current study of 76% SWs and 24% GSWs is
consistent with his findings.
We continue to base our management algorithms for traumatic
pleural collections on a policy of selective conservatism and this
approach appears to be supported by the current data with the
majority of penetrating thoracic trauma related injuries being
successfully managed conservatively [4,10–14]. However, approx-
imately 15% of all penetrating thoracic injuries should require
operative management and our operative rate is well below this
[10,11]. This almost certainly reflects our immature trauma system
with delayed evacuation of trauma victims being the norm. It is
feasible that if we improve our retrieval system, our rate of
operative exploration may well increase. From our mortuary data,
GSWs of the chest remain far more likely to be lethal than simple
SWs and our previous publications on this topic have consistently
shown that between 30 and 40% of victims of a GSW to the chest
will die on the scene [15,16]. On-going mortuary surveillance is
essential and we are continuing to develop strategies to improve
this.
Pulmonary GSWs do have a different spectrum of injury to SWs
[2]. SWs are more likely to cause a simple PTX, which is easily
(%)
[1] Oparah SS, Mandal AK. Penetrating gunshot wounds of the chest in civilian
43
practice: experience with 250 consecutive cases. Br J Surg 1978;65(January
16 (1)):45–8.
[2] Borja AR, Ransdell HT. Treatment of penetrating gunshot wounds of the chest.
20
Experience with one hundred forty-five cases. Am J Surg 1971;122(July
(1)):81–4.
[3] Muckart DJ, Meumann C, Botha JB. The changing pattern of penetrating torso
trauma in KwaZulu/Natal – a clinical and pathological review. S Afr Med J
1995;85(November (11)):1172–4.
[4] Clarke DL, Thomson SR, Madiba TE, Muckart DJ. Selective conservatism in
trauma management: a South African contribution. World J Surg 2005;29(Au-
gust (8)):962–5.
[5] Hegarty MM. A conservative approach to penetrating injuries of the chest.
Experience with 131 successive cases. Injury 1976;8(August (1)):53–9.
[6] Muckart DJ, Luvuno FM, Baker LW. Penetrating injuries of the pleural cavity.
Thorax 1984;39(October (10)):789–93.
[7] Madiba TE, Thomson SR, Mdlalose N. Penetrating chest injuries in the firearm
era. Injury 2001;32(January (1)):13–6.
[8] Laing GL, Skinner DL, Bruce JL, Aldous C, Oosthuizen GV, Clarke DL. Under-
standing the burden and outcome of trauma care drives a new trauma systems
model. World J Surg 2014 [Epub ahead of print].
[9] Solomon L, von Rahden RP, Allorto NL. Intra-operative cell salvage in South
Africa: feasible, beneficial and economical. S Afr Med J 2013;103(August
(10)):754–7. http://dx.doi.org/10.7196/samj.7355.
[10] Demetriades D, Rabinowitz B, Markides N. Indications for thoracotomy in stab
injuries of the chest: a prospective study of 543 patients. Br J Surg
1986;73:880–90.
[11] Demetriades D, Velmahos GC. Penetrating injuries of the chest: indications for
operation. Scand J Surg 2002;91(1):41–5.
[12] Norberg J, Nilsson T, Eriksson A, Hardcastle T. The costs of a bullet: inpatient
costs of firearm injuries in South Africa. S Afr Med J 2009;99(June (6)):442–4.
[13] Mattox KL, Allen MK. Systematic approach to pneumothorax, haemothorax,
pneumomediastinum and subcutaneous emphysema. Injury 1986;17(Sep-
tember (5)):309–12.
[14] Maritz D, Wallis L, Hardcastle T. Complications of tube thoracostomy for chest
trauma. S Afr Med J 2009;99(February (2)):114–7.
[15] Clarke DL, Gall TM, Thomson SR. Double jeopardy revisited: clinical decision
making in unstable patients with, thoraco-abdominal stab wounds and,
potential injuries in multiple body cavities. Injury 2011;42(May (5)):478–
81. http://dx.doi.org/10.1016/j.injury.2010.06.027.
[16] Clarke DL, Quazi MA, Reddy K, Thomson SR. Emergency operation for pene-
trating thoracic trauma in a metropolitan surgical service in South Africa. J
Thorac Cardiovasc Surg 2011;142(September (3)):563–8. http://dx.doi.org/
10.1016/j.jtcvs.2011.03.034.