Injury, Int. J.

Care Injured (2008) 39, 44—49

www.elsevier.com/locate/injury

The risk factors and management of

posttraumatic empyema in trauma patients
Sevval Eren a,*, Hidir Esme b, Abidin Sehitogullari c, Atilla Durkan a

a
Department of Thoracic Surgery, Dicle University, School of Medicine, 21280 Diyarbakir, Turkey
b
Department of Thoracic Surgery, Kocatepe University, School of Medicine, Afyon, Turkey
c
Department of Thoracic Surgery, General Hospital, Van, Turkey

Accepted 5 June 2007

KEYWORDS Summary
Trauma;
Tube thoracostomy; Background: Posttraumatic empyema increases patient morbidity, mortality and
Empyema; length of hospital stay, and the cost of treatment. The aim of this study was to
Risk factors identify the risk factors for posttraumatic empyema and to review our treatment
outcomes in patients with this condition.
Methods: A total of 2261 patients who were admitted with thoracic traumas and
underwent tube thoracostomy between January 1989 and January 2006 were inves-
tigated retrospectively. Posttraumatic empyema developed in 71 patients. Logistic
regression was used to assess the association between potential risk factors for
posttraumatic empyema. All values were expressed as the mean  S.D.
Results: Eight hundred and thirty-six (37%) of the patients had penetrating type
trauma, while 1425 (63%) had blunt type trauma. The rate of posttraumatic empyema
development was 3.1% for all patients. Pulmonary contusion was seen in 221 (9.8%)
patients and fractures of more than two ribs were seen in 191 (8.4%) patients. Tube
thoracostomy placement was performed in the emergency room in 1728 (76.4%)
patients, in the hospital ward in 197 (8.7%), in the intensive care unit in 182 (8.0%),
and in the operating room in 154 (6.8%). The duration of tube thoracostomy was
6.11  2.99 (1—21) days. Retained haemothorax was seen in 175 (7.7%) patients. The
mean lengths of hospital and intensive care unit stay were 6.42  3.45 and
2.36  2.66 days, respectively. The analysis showed that duration of tube thoracost-
omy (OR, 2.49, p < 0.001), length of intensive care unit stay (OR, 4.21, p < 0.001),
and presence of contusion (OR, 3.06, p < 0.001), retained haemothorax (OR, 5.55,
p < 0.001), and exploratory laparotomy (OR, 2.46, p < 0.001) were independent
predictors of posttraumatic empyema. The relative risk of posttraumatic empyema
was higher than 1 for each of the following risk factors: penetrating trauma (OR, 1.59,

* Corresponding author. Tel.: +90 412 2488001x4276; fax: +90 412 2488440.
E-mail address: [email protected] (S. Eren).

0020–1383/$ — see front matter # 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.injury.2007.06.001
Posttraumatic empyema in trauma patients 45

p = 0.055), associated injuries (OR, 1.12, p = 0.628) and fractures of more than two
ribs (OR, 1.60, p = 0.197).
Conclusion: Prolonged duration of tube thoracostomy and length of intensive care
unit stay, and the presence of contusion, laparotomy and retained haemothorax are
independent predictors of posttraumatic empyema. Use of prophylactic antibiotics
may be recommended in patients with these risk factors.
# 2007 Elsevier Ltd. All rights reserved.

Introduction Vital signs were taken immediately upon admis-

sion in the emergency room. After the initial phy-
Posttraumatic empyema (PTE) is a significant pro- sical examination and stabilisation of vital signs, a
blem in both blunt and penetrating chest injuries. CXR (chest X-ray) was obtained in all patients.
The incidence of PTE in patients who have sustained Thoracic injuries were diagnosed on CXR and thorax
injuries to the chest has been reported to range computerised tomography (CT) when necessary.
from 2% to 25%.1,5,7,9 Potential causes for PTE Indications for an emergency thorax CT in haemo-
include iatrogenic infection of the pleural space dynamically stable patients were as follows: a
during chest tube placement, direct infection widened mediastinum on thorax radiographs, multi-
resulting from penetrating injuries of the thoracic ple rib fractures, to be able to distinguish between
cavity, secondary infection of the pleural cavity pulmonary contusion with or without lacerations
from associated intra-abdominal organ injuries and pulmonary haematoma that might be undetect-
with diaphragmatic disruption, secondary infection able on CXR, and an unreliable physical and CXR
of undrained or inadequately drained haemotho- examination.
races, haematogenous or lymphatic spread of sub- The indication for tube placement was the pre-
diaphragmatic infection to the pleural space, and sence of pneumothorax and/or haemothorax. The
parapneumonic empyema resulting from posttrau- clinical course was evaluated based on changes
matic pneumonia, pulmonary contusion, or acute noted on CXR as well as relief of symptoms such
respiratory distress syndrome.16 as fever, dyspnoea, and tachypnea and change in
PTE causes an increase in the duration of hospi- character of the pleural fluid draining from the chest
talisation, in morbidity and mortality rates, and in tube. Antibiotics were not given routinely to all
the cost of treatment. Therefore, a good knowledge patients solely because of tube thoracostomy. Our
of the risk factors for the development of PTE is criteria for antibiotic prophylaxis included emer-
necessary and measures should be taken to prevent gency or urgent thoracotomy, soft-tissue destruc-
it. The role of prophylactic antibiotics is controver- tion of the chest wall by shotgun injuries, or
sial. In addition, risk factors that may independently associated open long bone fractures. The following
predict PTE in patients with tube thoracostomy patients were excluded from the study: those taking
have not been elucidated fully. The aim of this prophylactic antibiotherapy; those requiring thora-
study was to identify the risk factors for PTE and cotomy after tube thoracostomy for initial massive
to review our treatment outcomes in patients with haemothorax, continued bleeding or massive air
this condition. leak; those with a systemic disease such as diabetes
mellitus, malignancy, and liver disease; and those
taking prophylactic antibiotherapy for associated
Patients and methods injuries.
The diagnosis of empyema required one of the
Between January 1989 and January 2006, 2261 following criteria: (1) grossly purulent pleural fluid
patients who underwent tube thoracostomy for documented by thoracentesis; (2) positive pleural
thoracic trauma were treated in our Thoracic Sur- fluid culture or Gram stain; or (3) pleural fluid
gery clinics. The hospital records of these patients pH < 7.10, LDH > 1000 IU/L, or glucose level under
were reviewed retrospectively. Clinical data includ- 40 mg/dL. Once the diagnosis of PTE was achieved,
ing age, sex, type of injury, associated injuries, antibiotic treatment was initiated. The treatment
duration of tube thoracostomy, setting of tube thor- options are continued or repeated tube thoracost-
acostomy placement, length of hospital and inten- omy, intrapleural instillation of fibrinolytics, thor-
sive care unit stay, lung and chest wall injuries, acoscopy with the breakdown of adhesions or
presence of retained haemothorax and exploratory decortication, thoracotomy with the breakdown
laparotomy, the pleural cultures, presence of PTE, of adhesions and decortication, and open drainage
management and outcomes were recorded. procedures.
46 S. Eren et al.

The PTE continued to be treated with tube thor- (10.4%). The indication was not documented in six
acostomy and antibiotics as long as the chest tube patients. Tube thoracostomy placement was per-
continued draining pus and a thorax CT scan of the formed in the emergency room in 1728 (76.4%)
chest demonstrated that the chest tube was within patients, in the hospital ward in 197 (8.7%), in
the empyema cavity, there was no evidence of the intensive care unit in 182 (8.0%), and in the
multiple loculations or there was no contrast operating room in 154 (6.8%). The mean duration of
enhancement of the parietal pleura. Streptokinase tube thoracostomy was 6.11  2.99 (1—21) days.
was administered through a standard chest tube for Mean duration of tube placement was 7.52  3.69
fibrinolytic therapy. The dose, dwelling time of the days in patients with PTE and 6.07  2.95 days in
fibrinolytic agent, dosing interval, and number of those without.
treatments were determined by the surgeon respon- The presence of retained haemothorax was seen
sible for the patient’s care. in 175 (7.7%) patients. Retained haemothorax was
The indications for decortication were high-den- defined as pleural effusion significantly noted on
sity fluid or enhancement of the parietal pleura on chest X-ray despite tube thoracostomy in patients
CT scan, a multiloculated fluid collection, abscess with haemothorax/haemopneumothorax. All the
cavity within the posterior gutter of the chest, patients underwent ultrasound-guided tube thora-
continued drainage of pus through the chest tube costomy. Complete drainage was not achieved in 21
despite appropriate antibiotic therapy, and contin- out of the 175 patients and PTE developed. Among
ued fever and a homogeneous density within the the patients who developed PTE, fibrinolytic treat-
pleural cavity despite a functioning chest tube and ment was successful in 9 and decortications were
appropriate antibiotic therapy. successful in 12. The mean length of intensive care
All values were expressed as the mean  S.D. To unit stay was 2.36  2.66 (0—18) days. The mean
assess the association between potential risk factors length of hospital stay was 6.42  3.45 (4—52) days.
and PTE, odds ratios (OR) (confidence intervals, 95%) Forty-two of the 71 patients with PTE were suc-
were calculated using logistic regression. A p-value of cessfully treated with continued or repeated tube
<0.05 was considered statistically significant. thoracostomy alone. Seventeen patients were trea-
ted with continued tube thoracostomy and fibrino-
lytic therapy. Treatment failed in three of these
Results patients. Bleeding was observed in one of the
patients receiving fibrinolytic treatment, but this
The mean age of all patients was 37.2  20.6 (5—78) did not cause a decrease in haemoglobin values.
years, and 1560 (69%) of the patients were males. Fifteen of the 71 patients underwent decortication
The rate of PTE development was 4.0% (n = 34) in with thoracoscopy or thoracotomy. Thoracoscopy
patients with penetrating thoracic trauma, and 2.6% was performed in seven patients. Thoracoscopic
(n = 37) in those with blunt thoracic trauma. This decortication was unsuccessful in all patients in
rate was 3.1% (n = 71) for all patients. Firearm whom fibrinolytic therapy failed. These patients
wounds existed in 20 of the 34 (58.8%) patients with underwent thoracotomy. Decortication with thora-
penetrating thoracic trauma who developed PTE, cotomy was performed in 11 patients. Eight of these
while the remaining 14 (41.2%) were wounded with
sharp or piercing objects.
The most common associated injuries were inju- Table 1 Intrathoracic and thoracic wall injuries
ries of the abdomen (37.4%), followed by extremity Injury n/%
injuries (25.2%), injuries of the head (23.8%), inju- Haemothorax 1040/45.9
ries of the face (6.5%), and other injuries (1.95%). Haemopneumothorax 978/43.4
One thousand and nineteen patients (45.1%) had Pneumothorax 237/10.4
isolated thoracic injuries. Three hundred and four- Pulmonary contusion 221/9.8
teen patients (13.9%) underwent exploratory lapar- Intraparanchymal haematoma 29/1.3
otomy, 58 of which (18.4%) were negative. Table 1 Lung laceration 13/0.6
summarises the intrathoracic and thoracic wall inju- Tracheobronchial injury 7/0.3
ries. The most common intrathoracic injury was Less than three rib fractures 684/30.3
haemothorax, in 1040 patients (45.9%), and the More than three rib fractures 191/8.4
Subcutaneous emphysema 108/4.8
most common chest wall injury was rib fractures,
Flail chest 61/2.7
in 875 patients (38.7%).
Clavicular fracture 45/2.0
The indications for tube thoracostomy were hae- Sternal fracture 29/1.3
mothorax in 1040 patients (45.9%), haemopneu- Scapular fracture 20/0.9
mothorax in 978 (43.4%) and pneumothorax in 237
Posttraumatic empyema in trauma patients 47

Table 2 Logistic regression analysis for posttraumatic empyema rate (n = 71)

Risk factors Class PTE rate (%) Odds ratio Estimate CI 95% (L—U) p-Value
Type of injury Penetrating 34/836 (4) 1.590 (0.990—2.554) 0.055
Blunt 37/1425 (3)
Associated injuries (+) 41/1242 (3) 1.125 (0.698—1.816) 0.628
( ) 30/1019 (3)
Setting of tube placement Emergency 46/1728 (3) 0.556 (0.338—0.913) 0.071
Others 25/533 (5)
Contusion (+) 17/221 (8) 3.065 (1.744—5.386) 0.000
( ) 54/2040 (3)
More than two rib fractures (+) 9/191 (5) 1.602 (0.783—3.275) 0.197
( ) 62/2070 (3)
Retained haemothorax (+) 21/175 (12) 5.553 (3.251—9.484) 0.000
( ) 50/2086 (3)
Laparotomy (+) 27/463 (6) 2.469 (1.512—4.032) 0.000
( ) 44/1798 (3)
Tube duration >6 43/878 (9) 2.492 (1.526—4.042) 0.000
6 28/1383 (2)
Length of ICU stay >2 50/841 (6) 4.211 (2.511—7.062) 0.000
2 21/1420 (1)
Abbreviation–—PTE: posttraumatic empyema; CI: confidence interval; L—U: lower—upper; ICU: intensive care unit.

11 patients were successfully treated with decortica- tube thoracostomy for traumatic haemopneu-
tion. Of the remaining three patients, one recovered mothorax.16
after conversion to open thoracostomy, and the other In our 2261 patients who underwent tube thor-
two died. One patient died due to pulmonary embo- acostomy for thoracic trauma, we determined that
lism and one patient died due to cardiac failure. prolonged duration of tube thoracostomy more than
A pathogenic microorganism was grown in culture 6 days, prolonged intensive care unit stay more than
in 34 patients (47.8%) with PTE. The most commonly 2 days, and presence of contusion, retained hae-
identified pathogenic microorganisms were Staphy- mothorax and laparotomy were independent pre-
lococcus aureus, in 14 patients, and gram negative dictors of PTE. We recommend use of prophylactic
bacilli, in nine patients. Anaerobic bacteria were antibiotics only in patients with these risk factors,
identified in four patients, and mixed bacterial and we think that routine use of prophylactic anti-
populations in the remaining seven. biotics in patients with tube thoracostomy for
The logistic regression analysis showed that dura- thorax traumas is unnecessary. On the other hand,
tion of tube thoracostomy, length of intensive care in our study, only 3.1% of patients with chest trauma
unit stay, and presence of contusion, retained hae- developed PTE. Therefore, there appears to be no
mothorax, and exploratory laparotomy were inde- need to use antibiotics routinely in patients with a
pendent predictors of PTE (Table 2). chest tube.
The rate of PTE development following penetrat-
ing wounds varies between 1.7% and 25% in the
Discussion literature, whereas this rate is reported to be
between 1.6% and 17% in blunt thoracic trau-
The use of prophylactic antibiotics for the preven- mas.5,9,10,13,26 This rate generally varies between
tion of PTE after tube thoracostomy remains a 2% and 25% for the thorax injuries as a whole.1,5,7,9
controversial issue in the trauma literature. While The rates of PTE in our patients were 2.6% for blunt
a number of studies3,4,11,15,18,22,25 show favour- traumas, 4.0% for penetrating traumas, and 3.1% for
able effects of prophylactic antibiotics in prevent- all thoracic traumas. PTE resulting from penetrating
ing PTE, several reports12,14,17,20 have shown no trauma, particularly due to gunshot wounds, was
benefit. The EAST Practice Management Guide- encountered more frequently (20 of 34 patients with
lines Work Group concluded that there were insuf- penetrating thoracic trauma). In our study, pene-
ficient data available to recommend routine use trating trauma was not an independent predictor of
of prophylactic antibiotics in the management of PTE in the logistic regression analysis.
48 S. Eren et al.

In Mandal’s landmark review of 5474 patients The three patients in whom thoracoscopy failed had
with tube thoracostomy for thoracic trauma, the all received fibrinolytic treatment before. During
only risk factor identified for the development of the operation, it was observed that the three
PTE was retained haemothorax.18 The authors con- patients in whom fibrinolytic therapy failed had
cluded that adequate drainage of the pleural space a thick pleural peel that was not evident on CT
was imperative, and, when achieved, empyema scans. We suggest that CT evidence of a thickened
rates were low, around 1.6%. Eddy et al.9 demon- pleural peel uniformly predicts failure of fibrinolytic
strated similar results for retained haemothorax and therapy and is an indication for primary operative
showed that chest tubes that retained for prolonged management.
periods also resulted in increased rates of PTE.
In the present study, retained haemothorax on
immediate follow-up chest radiograph occurred in Conclusion
175 (7.7%) patients and was associated with PTE in
21 patients (12%). We found that tubes remaining for We determined that prolonged duration of tube
longer durations were associated with greater than thoracostomy and length of intensive care unit stay,
fivefold increased incidence of PTE. In patients with and presence of contusion, laparatomy and retained
severe blunt trauma, massive contusion and ribs haemothorax were independent predictors of PTE.
fractures are frequently encountered. These condi- Use of prophylactic antibiotics may be recommended
tions lead to conditions predisposing to PTE such as only in patients with these risk factors. Because of the
atelectasia, pneumonia, and haematoma. In thor- low incidence of posttraumatic empyema thoracis,
acoabdominal injuries, enteric bacteria can be we do not recommend routine antibiotic prophylaxis
translocated to the thorax. We found that the pre- for all trauma patients who undergo tube thoracost-
sence of contusion and laparotomy was associated omy. Fibrinolytic therapy and thoracoscopy seem to
with greater than threefold and twofold increases in be effective methods for definitive treatment in
the incidence of PTE, respectively. selected patients if performed early.
In our study, 42 of the 71 patients with PTE were
treated with continued or repeated tube thoracost-
omy alone. In all patients who underwent this References
procedure the PTE resolved. When fibrinolytic treat-
ment was used with continued tube thoracostomy, 1. Arom KV, Grover FL, Richardson JD, Trinkle JK. Post-trau-
PTE resolved in 14 of 17 patients. Treatment failed matic empyema. Ann Thorac Surg 1977;23:254—8.
in three of these patients. Intrapleural fibrinolytics, 2. Bouros D, Schiza S, Tzanakis N, et al. Intrapleural urokinase
if used early in the fibrinopurulent stage of a para- versus normal saline in the treatment of complicated para-
pneumonic effusions and empyema. A randomized, double-
pneumonic effusion before extensive collagen is blind study. Am J Respir Crit Care Med 1999;159:37—42.
deposited in the pleural space, decrease the rate 3. Brunner RG, Vinsant GO, Alexander RH, et al. The role of
of surgical interventions (VATS or open decortica- antibiotic therapy in prevention of empyema in patients with
tion) and the length of hospital stay with minor an isolated chest injury (ISS 9-10): a prospective study. J
associated morbidity. However, this therapeutic Trauma 1990;30:1148—54.
4. Cant PJ, Smart DO. Antibiotic prophylaxis is indicated for
modality to help to break the loculations by virtue chest stab wounds requiring closed tube thoracostomy. Br J
of its fibrinolytic property is still controversial.21 Surg 1993;80:464—6.
Prospective studies during the last decade have 5. Caplan ES, Hoyt NJ, Rodriguez A, Cowley RA. Empyema
come to contradictory conclusions.2,6,8,19 occurring in the multiple injured traumatized patient. J
Trauma 1984;24:785—9.
Fifteen of the 71 patients with PTE required
6. Chin NK, Lim TK. Controlled trial of intrapleural streptoki-
decortication. Thoracoscopy was performed in nase in the treatment of pleural empyema and complicated
seven patients. In three of the seven patients who parapneumonic effusions. Chest 1997;111:275—9.
underwent thoracoscopy, it was converted to thor- 7. Coselli JS, Mattox KL, Beall Jr AC. Reevaluation of early
acotomy. With appropriate patient selection, thor- evacuation of clotted hemothorax. Am J Surg 1984;148:
acoscopy can effectively treat empyema thoracis by 786—90.
8. Davies RJ, Traill ZC, Gleeson FV. Randomised controlled trial
lysing adhesions, draining the abscess cavity, and of intrapleural streptokinase in community acquired pleural
decorticating the pleural peel. Thoracoscopy for infection. Thorax 1997;52:416—21.
PTE will be particularly helpful when exudative 9. Eddy AC, Luna GK, Copass M. Empyema thoracis in patients
pleural fluid is noted with numerous adhesions undergoing emergent closed tube thoracostomy for thoracic
and loculations, in the presence or absence of air trauma. Am J Surg 1989;157:494—7.
10. Fallon Jr WF. Post-traumatic empyema. J Am Coll Surg
leak. When there is a thick peel with significant lung 1994;179:483—92.
entrapment, however, thoracoscopy may be less 11. Gonzalez RP, Holevar MR. Role of prophylactic antibiotics for
than ideal compared with open thoracotomy.18,23,24 tube thoracotomy in chest trauma. Am Surg 1998;64:617—21.
Posttraumatic empyema in trauma patients 49

12. Grover FL, Richardson JD, Fewel BA, et al. Prophylactic 20. Maxwell RA, Campbell DJ, Fabian TC, et al. Use of pre-
antibiotics in the treatment of penetrating chest wounds. sumptive antibiotics following tube thoracostomy for trau-
J Thorac Cardiovasc Surg 1977;74:528—36. matic hemopneumothorax in the prevention of empyema
13. Heniford BT, Carrillo EH, Spain DA, et al. The role of thor- and pneumonia-a multi-center trial. J Trauma 2004;57:
acoscopy in the management of retained thoracic collections 742—8.
after trauma. Ann Thorac Surg 1997;63:940—3. 21. Misthos P, Sepsas E, Konstantinou M, et al. Early use of
14. LeBlanc KA, Tucker WY. Prophylactic antibiotics and closed intrapleural fibrinolytics in the management of postpneumo-
tube thoracostomy. Surg Gynecol Obstet 1985;160:259—63. nic empyema. A prospective study. Eur J Cardiothorac Surg
15. LoCurto Jr JJ, Tischler CD, Swan KG, et al. Tube thoracost- 2005;28:599—603.
omy–—antibiotics or not? J Trauma 1986;26:1067—72. 22. Nichols RL, Smith JW, Muzik AC, et al. Preventive antibiotic
16. Luchette FA, Barrie PS, Oswanski MF, et al. Practice manage- usage in traumatic thoracic injuries requiring closed tube
ment guidelines for prophylactic antibiotic use in tube thoracostomy. Chest 1994;106:1493—8.
thoracostomy for traumatic hemopneumothorax: the EAST 23. O’Brien J, Cohen M, Solit R, et al. Thoracoscopic drainage and
Practice Management Guidelines Work Group. Eastern Asso- decortication as definitive treatment for empyema following
ciation for Trauma. J Trauma 2000;48:753—7. penetrating chest injury. J Trauma 1994;36:536—9.
17. Mandal AK, Montano J, Thadepalli H. Prophylactic antibiotics 24. Ridley PD, Braimbridge MV. Thoracoscopic debridement and
and no antibiotics compared in penetrating chest trauma. J pleural irrigation in the management of empyema thoracis.
Trauma 1985;25:639—43. Ann Thorac Surg 1991;51:461—4.
18. Mandal AK, Thadepalli H, Chettipalli U. Posttraumatic 25. Stone HH, Panagiotis NS, Hooper CA. Cefamandole for pro-
empyema thoracis: a 24 year experience at a major trauma phylaxis against infection in closed tube thoracostomy. J
center. J Trauma 1997;43:764—71. Trauma 1981;21:975—7.
19. Maskell NA, Davies CW, Nunn AJ, et al. U.K. controlled trial of 26. Wilson JM, Boren Jr CH, Peterson SR, Thomas AN. Traumatic
intrapleural streptokinase for pleural infection. N Engl J Med hemothorax: is decortication necessary? J Thorac Cardiovasc
2005;352:865—74. Surg 1979;77:489—95.