European Journal of Cardio-thoracic Surgery 23 (2003) 473–476

www.elsevier.com/locate/ejcts

Delayed pericardial effusion following stab wounds to the chest

David G. Harris*, Jacques T. Janson, Jacques Van Wyk, Johann Pretorius, Gawie J. Rossouw

Downloaded from https://academic.oup.com/ejcts/article/23/4/473/382104 by guest on 03 November 2021

Department of Cardiothoracic Surgery, Tygerberg Hospital, University of Stellenbosch, Cape Town, South Africa
Received 30 September 2002; received in revised form 16 December 2002; accepted 29 December 2002

Abstract
Introduction: Delayed pericardial effusion following penetrating cardiac trauma has not been commonly reported, and the exact incidence
remains unknown. It was more common before 1960, when pericardiocentesis was still a popular treatment for stable patients presenting with
a stab wound to the heart. Material and methods: During an 8-year period, 24 patients were diagnosed with delayed pericardial effusions
following a recent stab wound over the chest. Nine patients had been initially treated at our trauma unit, and the remaining 15 patients were
referred by a peripheral clinic. Results: Diagnosis was confirmed by cardiac ultrasound or echocardiogram. Sixteen patients were adequately
treated by subxiphoid drainage. Sternotomy was performed in five patients, left thoracotomy in two and right thoracotomy in one patient. No
actively bleeding injuries were found. Three patients had active infection in the pericardial space. Fever, pleural effusions and ascites were
common associated findings. Additional procedures performed included laparotomy for acute abdominal pain in two patients (both negative),
and simultaneous drainage of a pleural empyema. Two patients with staphylococcal pericardial infections required subsequent
pericardiectomy. Summary: The diagnosis of a penetrating cardiac patient may be missed in a stable patient, and patients may present with
delayed pericardial effusions and tamponade. Post pericardiotomy syndrome may be the most common cause of delayed pericardial effusion,
followed by sepsis. Subxiphoid pericardial window is an adequate form of treatment. Recent literature reveals that occult cardiac injury is not
uncommon, thus a case should be made to actively investigate all patients with precordial stab wounds with cardiac ultrasound or
echocardiogram.
q 2003 Elsevier Science B.V. All rights reserved.
Keywords: Delayed pericardial effusion; Cardiac ultrasound

1. Introduction 2. Patients and methods

Acute pericardial tamponade following penetrating
cardiac trauma is common. Delayed pericardial effusion,
however, has rarely been described. It has not been
commonly reported in the literature since 1960, when
pericardiocentesis was still a common treatment for stable
patients presenting with acute cardiac tamponade following
a stab. Subsequently, penetrating heart injuries have been
treated by prompt surgical exploration and repair.
Over an 8-year period, 24 patients were presented to our
hospital with delayed pericardial effusion. We discuss the
clinical features, diagnosis and treatment of these cases.
* Corresponding author. Tel.: þ 27-21-9762347; fax: þ 27-21-9385952.
E-mail address: drdharris@yahoo.co.uk (D.G. Harris).
1010-7940/03/$ - see front matter q 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S1010-7940(03)00006-X
There were a total of 24 patients from January 1994 to
December 2001. There were 23 males and one female, with
an average age of 28 years (14 – 53 years). Twelve patients
were initially treated for the stab wound at a peripheral
hospital or clinic; seven of these had intercostal drains
inserted, and the rest were sutured and discharged. Three
patients were referred from the periphery with no mention
whether they had been previously seen. Nine patients were
originally treated at our trauma unit. They all had a single
parasternal or precordial wound, and all were stable, except
for one patient who had a blood pressure of 80/40 mmHg for
3 h. Three of these nine patients required initial intercostal
drain insertion for haemothorax. The remaining six were
sutured and discharged.
Four patients were admitted to our medical ward with a
diagnosis of pericardial effusion before the history of trauma
became evident. Patients were examined clinically, and the
diagnosis was usually confirmed by two-dimensional
474 D.G. Harris et al. / European Journal of Cardio-thoracic Surgery 23 (2003) 473–476

ultrasound, or preferably by echocardiogram, if it was
immediately available. Surgical approach was left to the
surgeon’s discretion.
3. Results
The most common clinical findings were distended neck
veins, dyspnoea, pleural effusion and other features of right
heart failure (Table 1). Fever occurred in eight patients
(33%). A globular heart on chest X-ray was documented in
20 patients, and one patient had a wide mediastinum. The
Two litres of ascitic fluid was drained, and no
abdominal injury was found. Staphylococcus aureus
was cultured from the pericardial fluid. Two weeks
later, the patient underwent sternotomy and pericar-
diectomy for severe constrictive pericarditis.
(ii) Patient 12 underwent subxiphoid drainage and had a
positive culture of Staphylococcus aureus from the
pericardial fluid. One week later a right thoracotomy
and drainage of empyema was performed. Four months
later, the patient underwent sternotomy and pericar-
diectomy for severe constrictive pericarditis.
(iii) Patient 14 underwent laparotomy for an acute abdomen

Downloaded from https://academic.oup.com/ejcts/article/23/4/473/382104 by guest on 03 November 2021

age of the stab varied from 3 to 33 days. Diagnosis was at the same time as the subxiphoid drainage. Hepato-
confirmed by cardiac ultrasound in 16 patients, echocardio- megaly and ascites were the only abnormal findings.
graphy in seven and clinically in one patient. (iv) Patient 21 had a sternotomy for suspected acute
Drainage was performed by subxiphoid window in 16 pericardial tamponade. This patient had been
patients, sternotomy in five, left thoracotomy in two and stabbed twice, 2 weeks earlier and on the day of
right thoracotomy in one case. Except for patient 21, there admission. Cardiac ultrasound revealed a pericardial
was no specific reason for sternotomy, but surgeon’s effusion and the patient was taken to the operating
preference. Not one actively bleeding wound was found. room with a suspected acute heart injury. Blood
Drainage via the subxiphoid approach was adequate stained fluid and old clots were found in the
when performed. A pericardial window was made, draining pericardial space and a healing scar was noticed
blood clots and blood stained fluid. The whole heart surface on the right ventricle. The new stab did not
could not be examined for injuries with this approach, but it penetrate the pericardium and the pericardial tam-
allowed some visualization of the right ventricle, and ponade was believed to be caused by the first stab
adequate drainage of the pericardial space was possible. wound.
Conversion from subxiphoid to a sternotomy was never
found to be necessary. Three patients had positive cultures There was no mortality. Two patients required inotropic
from pericardial fluid (Staphylococcus aureus in two support for 2 days and one patient required ventilation for 2
patients and beta-haemolytic streptococcus in one). One days. Except for the two patients with constrictive
patient had a co-existing empyema and the pericardial pericarditis (following Staphylococcus aureus infection),
effusion and empyema were drained simultaneously by right all the patients had uncomplicated postoperative courses.
thoracotomy. Two patients had moderate mitral regurgitation at later
Additional procedures were performed as follows: (Table follow-up as a result of leaflet prolapse.
2) During the year 2001, 1066 patients were treated at our
trauma unit for a stab wound in the chest. Seven hundred
and eighty-nine required chest drain insertion for hae-
(i) Patient 1 underwent laparotomy at the same time as the mothorax or pneumothorax, 249 had only flesh wounds and
subxiphoid drainage, for suspected subphrenic abscess. 61 had surgery for a penetrating cardiac injury.
Table 1
Clinical findings
4. Discussion
Distended neck veins 20 (83%)
Dyspnoea 13 (54%) Delayed pericardial effusion/haemopericardium after
Pleural effusion 10 (41%)
chest trauma is uncommon, as suspected cardiac injuries
Fever 8 (33%)
Hepatomegaly 8 (33%) are now-a-days treated by prompt surgical exploration. In
Abdominal pain 7 (29%) our unit, cardiac ultrasound has been useful in diagnosing
Oedema 6 (25%) abnormal pericardial fluid in doubtful cases [1,2].
Pleuritic pain 4 (17%) Some patients presenting with penetrating cardiac
Ascites 3 (12%)
Soft heart sounds 3 (12%)
injuries may be completely stable, and the diagnosis can
Precordial pain 3 (12%) be missed [3]. These patients do not present with the
Hypotension 2 (8%) classical findings of raised venous pressure, hypotension
Friction rub 2 (8%) and muffled heart sounds. The chest X-ray does not always
Systolic murmur 2 (8%) reveal an enlarged heart shadow. In our series of 191
EKG signs of pericarditis 1 (4%)
Anaemia 1 (4%)
patients with penetrating cardiac injuries, 63 patients (33%)
were completely stable after resuscitation with 2 l of fluid.
 D.G. Harris et al. / European Journal of Cardio-thoracic Surgery 23 (2003) 473–476 475

Table 2
Perioperative dataa

Patient Age of stab (days) Position of stab Size of effusion (mm) Surgical approach Amount of fluid (ml) Quality of fluid

1 8 Eighth left rib space ‘Large’ Subxiphoid ‘Large’ N/S

2 16 Left parasternal 30 Minithoracotomy 1000 B/S
3 4 Mid sternum 10 Sternotomy 300 Straw coloured
4 14 Left parasternal 30 Subxiphoid 800 B/S
5 29 Not documented ‘Large’ Subxiphoid ‘Massive’ B/S
6 8 Left parasternal 60 Subxiphoid 500 B/S
7 31 Right parasternal 70 Subxiphoid 350 B/S
8 12 Left parasternal 20 Left thoracotomy 200 B/S and clots
9 14 Left parasternal 18 Sternotomy 1000 B/S and clots

Downloaded from https://academic.oup.com/ejcts/article/23/4/473/382104 by guest on 03 November 2021

10 10 Left parasternal 20 Subxiphoid 500 Old blood
11 10 Multiple (5) anterior ‘Large’ Subxiphoid 300 B/S
12 15 Right anterior chest No ultrasound Subxiphoid 500 Turbid
13 7 Left parasternal ‘Large’ Subxiphoid 500 B/S
14 5 Right parasternal ‘Large’ Subxiphoid 200 Straw coloured
15 18 Xiphoid area 35 Sternotomy 500 B/S
16 16 Left parasternal 40 Subxiphoid 1200 B/S
17 22 Right parasternal ‘Large’ Subxiphoid 600 B/S
18 5 Right parasternal ‘Large’ Right thoracotomy 300 B/S
19 27 Left parasternal 32 Subxiphoid 100 Old blood
20 30 Left parasternal 15 Subxiphoid 200 B/S
21 14 Left parasternal £ 2 20 Sternotomy 300 B/S and clots
22 8 Left parasternal 15 Subxiphoid 350 B/S
23 15 Left parasternal 15 Subxiphoid 500 B/S
24 3 Left parasternal 30 Sternotomy 300 B/S and clots
a
N/S, not specified; B/S, bloodstained.

Fifty-six of these 63 patients (87%) had the diagnosis
confirmed by cardiac ultrasound. It is interesting to note that
during surgery, 27 of the 63 patients (43%) were found to
have lacerations which had sealed off [2]. It is in these
patients where the diagnosis may be missed during the
initial presentation, if ultrasound or echocardiogram is not
performed.
The most common clinical presentation of the patients
presenting with delayed pericardial effusion was dyspnoea
and distended neck veins, as well as pleural effusions,
ascites and fever (Table 1). Only two patients (8%)
presented with hypotension. Hepatomegaly and abdominal
pain were present in about one-third of the patients, and as a
result, laparotomy was performed in two of these patients.
The clinical presentation may be confusing in patients who
have an epigastric stab wound.
Diagnosis was confirmed by cardiac ultrasound in most
patients as it is used regularly in the acute setting in our
institution, and has been shown to be reliable, in the absence
of haemothorax [1,2]. Pericardiocentesis is not used in our
trauma unit in the acute setting for diagnosing cardiac
trauma, as it has been shown to be inaccurate [2], and this
may explain why this was not performed as an initial
diagnostic procedure.
Drainage via subxiphoid window was adequate in our
experience. Treatment by aspiration alone has been shown
to be inadequate in the acute setting, as recurrent pericardial
effusions may occur, requiring numerous aspirations [4,5].
Some patients may have blood clots in the pericardium
(Table 2), which may not be amenable to aspiration.
However, the rest could possibly have been treated by
aspiration and the insertion of an indwelling drainage
catheter, as we have not demonstrated any actively bleeding
lacerations. Thoracoscopy has been useful in stable patients
with acute penetrating cardiac wounds [6], and could be
another option in patients with delayed tamponade.
Sternotomy was performed in three of our patients as a
result of surgeon preference early on in our experience,
before it became evident that actively bleeding wounds
were never found. In two cases, sternotomy was performed
for a suspected acute injury. Sternotomy should not be
necessary as the initial management, and in fact, should be
avoided if infection is suspected. In the presence of
pyogenic infection within the pericardial and/or pleural
spaces, thoracoscopy or thoracotomy should be a better
approach. If pus is found within the pericardial space, it is
possibly best to perform pericardiectomy via thoracotomy
during the same setting, because of the risk of constrictive
pericarditis developing. Alternatively, these patients should
be carefully followed up for constrictive pericarditis.
Secondary bleeding, clot lysis with hyperosmosis of the
clot with fluid accumulation into the pericardium and post-
pericardiotomy syndrome are all possible causes of delayed
pericardial effusion. When aspiration was the favoured
treatment of stable patients with acute haemopericardium,
delayed haemopericardium was a common finding [4,7,8].
Some of these patients had pain, fever and residual
pericardial effusions for up to 30 weeks [4]. Two of these
476 D.G. Harris et al. / European Journal of Cardio-thoracic Surgery 23 (2003) 473–476
patients responded well to corticosteroids, and it was noted
that bilateral pleural effusions were common, probably the
result of hypersensitivity to blood in the pericardial space.
Two patients developed constrictive pericarditis [4].
In our series, pleural effusions were common, occurring
in ten of 24 patients (41.6%). These usually contained clear
fluid in longstanding cases, and were associated with
ascites. In addition, no patient had actively bleeding
lacerations, and this strengthens the case for post-pericar-
diotomy syndrome being the principle cause of delayed
pericardial effusion. Three cases of infection were proven to
contribute to or cause the effusion.
A case should be made to actively investigate all patients
initially presenting with penetrating precordial wounds. In
one study, echocardiogram was performed in 31 stable
patients with precordial injuries. There were 16 positive
echocardiograms, and all underwent subxiphoid explora-
tion, which yielded 12 positive and four negative results.
Three of the positive explorations demonstrated only
pericardial injuries [9]. Another group performed subxi-
phoid exploration in all patients with juxtacardiac injuries,
and revealed an occult injury rate in 12 of 51 stable patients
(17.6%) [10]. Some of the cardiac injuries were not actively
bleeding.
The largest review of patients with delayed pericardial
effusions since 1950 revealed eight patients [5]. Four were
treated by sternotomy, two by thoracotomy, one by
pericardiocentesis (five aspirations) and one was diagnosed
post-mortem. Since then there have been isolated reports in
the literature [11 – 14].
Missed injuries leading to delayed haemopericardium
with tamponade or death may be more common than
initially thought. One report comparing hospital and post-
mortem records of stab heart victims revealed that 70 of
1198 patients (5.8%) reached the hospital alive. However, a
further seven patients who had been admitted to hospital
with a chest wound and discharged, were amongst the group
of patients identified in the mortuary [15]. This reflects a
missed stab heart rate of 9% (seven of 77 patients), and this
is further evidence that all asymptomatic patients with
precordial wounds should undergo cardiac ultrasound or
echocardiogram to exclude a haemopericardium.
5. Conclusion
Delayed pericardial effusion/haemopericardium follow-
ing a precordial stab wound has not been commonly
described, but has been identified at our hospital where a
large volume of stab wounds are treated. An average of
three cases per year were diagnosed at our hospital, where
about 1000 cases of stab wounds to the chest are managed
annually. Missed stab wounds to the heart may occur if
routine cardiac ultrasound or echography is not performed
in all stable patients presenting with a stab wound in the
precordial area. Treatment by subxiphoid pericardial
window is adequate. Severe abdominal pain and ascites
may be an associated finding, leading to laparotomy.
Patients with septic pericarditis should be followed up for
the possible development of constrictive pericarditis at a
later stage, or could be initially drained by left thoracotomy
with pericardiectomy.
Downloaded from https://academic.oup.com/ejcts/article/23/4/473/382104 by guest on 03 November 2021
References
[1] Harris DG, Papagiannopoulos KA, Pretorius J, Van Rooyen T,
Rossouw GJ. Current evaluation of cardiac stab wounds. Ann Thorac
Surg 1999;68:2119– 22.
[2] Harris DG, Bleeker CP, Pretorius J, Rossouw GJ. Penetrating cardiac
injuries: current evaluation and management of the stable patient. S
Afr J Surg 2001;39:90–4.
[3] Major-Davies JA, D’Egidio A, Schein M. ‘Missed’ stabbed hearts –
pitfalls in the diagnosis of penetrating cardiac injuries. S Afr J Surg
1992;30:18–19.
[4] Tabatznik B, Isaacs JP. Postpericardiotomy syndrome following
traumatic hemopericardium. Am J Cardiol 1961;7:83–96.
[5] Aaland MO, Sherman RT. Delayed pericardial tamponade in
penetrating chest trauma: case report. J Trauma 1991;31(11):1563–5.
[6] Hoff WS, McMahon DJ, Schwab CW, Sing RF. Thoracoscopic
pericardial window and penetrating cardiac trauma. J Trauma 1997;
43:561.
[7] Heller RF, Rahimtoola SH, Ehsani A, Johnson S, Boyd DR, Tatooles
CJ, Loeb HS, Rosen KR. Cardiac complications: results of penetrating
chest wounds involving the heart. Arch Int Med 1974;135:491–6.
[8] Sugg WL, Rea WJ, Ecker RR, Web WR, Rose EF, Shaw RR.
Penetrating wounds of the heart: an analysis of 459 cases. J Thorac
Cardiovasc Surg 1968;56:531–45.
[9] Nagy KK, Lohmann C, Kim DO, Barrett J. Role of echocardiography
in the diagnosis of occult penetrating cardiac injury. J Trauma 1995;
38:859–62.
[10] Duncan AO, Scalea TM, Sclafani SJA, Phillips TF, Bryan D, Atweh
NA, Vieux EE. Evaluation of occult cardiac injuries using subxiphoid
pericardial window. J Trauma 1989;29:955.
[11] Hasegawa J, Noguchi N, Yamasaki J, Kotake H, Mashiba H, Sasaki S,
Mori T. Delayed cardiac tamponade and hemothorax induced by an
acupuncture needle. Cardiology 1991;78(1):58–63.
[12] Klinkenberg TJ, Kaan GL, Lacquet LK. Delayed sequelae of
penetrating chest trauma: a plea for early sternotomy. J Cardiovasc
Surg (Torino) 1994;35:173 –5.
[13] Raney JL, Kennedy ES. Delayed cardiac tamponade following a stab
wound: a case report. J Ark Med Soc 1997;93(12):589– 91.
[14] Mechem CC, Alam GA. Delayed cardiac tamponade in a patient with
penetrating chest trauma. J Emerg Med 1997;15(1):31–3.
[15] Campbell NC, Thomson SR, Muckart DJ, Meumann CM, Van-
Middelkoop I, Botha JB. Review of 1198 cases of penetrating cardiac
trauma. Br J Surg 1997;84:1737–40.