Original Contribution

The effect of different relieving methods on the outcome of out-of-hospital cardiac arrest patients with nontraumatic hemopericardium in the ED

Ruei-Fang Wang MD^a, Chun-Chieh Chao MD^b, Tzong-Luen Wang MD, PhD^a,c, Kuo-Chih Chen MD^a, Chee-Fah Chong MD, MS^a,c, Kuo-Hung Huang MD^d,

Chien-Chih Chen MD, MS^a,c,?

^aEmergency Department, Shin-Kong Wu Ho-Su Memorial Hospital, Taipei 111, Taiwan ^bDepartment of Emergency Medicine, Taipei City Hospital, Taipei 103, Taiwan ^cSchool of Medicine, Fu Jen Catholic University, Taipei 242, Taiwan

^dDepartment of Surgery, Taipei Veterans General Hospital, Taipei 112, Taiwan

Received 25 May 2007; revised 20 June 2007; accepted 10 July 2007

Abstract

Aims: This study aimed to assess the impact of different methods of draining nontraumatic hemopericardium on outcome from patients with out-of-hospital cardiac arrest (OHCA), identify independent predictors of return of spontaneous circulation (ROSC), and examine the ineffective rate of decompression based on subxiphoid pericardiotomy (SP) and percutaneous pericardial catheter drainage (PCD).

Methods: Adult patients with OHCA who presented to the ED between May 1, 2000, and October 30, 2006, with moderate to massive nontraumatic hemopericardium were recruited and stratified into 4 groups according to the relieving methods of hemopericardium. Charts were reviewed for various demographic data, resuscitation records, management, and outcome. Patient outcome was recorded as survival to hospital discharge and ROSC, as primary end points. Effective decompression was recorded as a secondary end point. We compared the outcome between the groups.

Results: A total of 1491 OHCA resuscitation records were prospective collected. There were 23 OHCA patients with moderate to massive nontraumatic hemopericardium. The overall ROSC rate was 39.1% (9/23). There was a clear difference in the ROSC rate between 4 groups (P b .05). The overall rate of survival to hospital discharge was 4.3% (1/23). There was no significant difference in the rate of survival to hospital discharge between the groups. Relieving methods was an independent predictor of ROSC in OHCA patients with nontraumatic hemopericardium (odds ratio, 0.17; 95% confidence interval, 0.4-0.70). There was a significant statistical difference in adequate relief of hemopericardium based on SP and PCD (P b .01).

* Corresponding author. Emergency Department, Shin-Kong Wu Ho-Su Memorial Hospital, Taipei City 111, Taiwan (ROC). Tel.: +886 2 28332211×2087; fax: +886 2 28353547.

E-mail address: timshel@ms67.hinet.net (C.-C. Chen).

0735-6757/$ – see front matter (C) 2008 doi:10.1016/j.ajem.2007.07.010

Conclusion: The early effective decompression method is associated with an increased rate of ROSC for OHCA patients with nontraumatic hemopericardium. Subxiphoid pericardiotomy has a better effective decompression of hemopericardium than PCD.

(C) 2008

Introduction

The hemodynamic spectrum of pericardial effusion ranges from asymptomatic, instability, life-threatening to death [1]. Development of acute cardiac tamponade is a life- threatening emergency requiring prompt treatment. Immedi- ately and effectively decompression is the mainstay of treatment. Cardiac tamponade could be relieved by pericar- diocentesis, percutaneous pericardial catheter drainage (PCD), conventional subxiphoid pericardiostomy, or sub- xiphoid pericardiotomy (SP). There were many studies discussing the drainage strategies for cardiac tamponade in aLive patients [2-5], but only a few concerning those without vital signs, for example, out-of-hospital cardiac arrest (OHCA) patients [6] or patients without palpable pulse [5,7]. It is a great issue in the ED. Furthermore, hemopericardium has increasing possibility of blood clot formation in the pericardial space as time goes by, which could make the drainage difficult, which is distinct from the other kind of pericardial effusion. Pericardial catheter drainage is not always effective in the case of hemoper- icardium. In the study by Kurimoto et al [6], SP is safe, effective, and could be performed quickly in an emergency situation in patients with cardiac tamponade because of acute hemopericardium. To our knowledge, no previous study has evaluated the results of SP in OHCA patients with nontraumatic hemopericardium. There was no study either comparing the results of different relieving methods of hemopericardium in OHCA patients with nontraumatic hemopericardium. In this study, we grouped OHCA patients with moderate to massive nontraumatic hemopericardium by relieving methods of hemopericardium to assess the impact of different methods of draining nontraumatic hemopericar- dium on outcome from the patients with OHCA, identified the independent predictors of the return of spontaneous circulation (ROSC), and examined the ineffective rate of decompression based on SP and PCD.

Material and methods

Study population and study design

The study was conducted in the ED of Shin-Kong Wu Ho- Su Memorial Hospital between May 1, 2000, and October 30, 2006. Shin-Kong Wu Ho-Su Memorial Hospital is a university teaching hospital with 921 beds and is one of 2 hospital centers in northern Taipei City in Taiwan. The ED has an adult intensive care unit bED capacity of 60 and

approximately 75000 ED visits annually. All adult patients (age N18 years) with OHCA who presented to the ED during this period with moderate to massive hemopericardium confirmed by ultrasound-guided pericardial diagnostic tap- ping were eligible for inclusion into the study. Patients were excluded from the study if they had a small amount of pericardial effusion, history of malignancy, previously documented pericardial effusion, trauma, or signed “do not attempt resuscitation” (DNAR). The study protocol was approved by our institutional review board.

Data collection and definitions

We have developed a form of resuscitation record since 2000, and nurses should well know about the fillings if they participated in the resuscitation. All patients with OHCA were enrolled prospectively in a database using a standard collection tool that was consistent with the Utstein criteria. The resuscitation records consist of 2 parts: one is filled by the emergency medical technicians and the other is filled by a nurse during the resuscitation period. All OHCA patients had been undergoing cardiopulmonary resuscitation with tra- cheal intubation according to the advanced cardiac life support guideline before or during pericardial decompres- sion. During pericardial decompression, we continuously performed chest compression as much as possible to minimize the interruption of resuscitation. Ultrasound was used to determine pericardial fluid, and hemopericardium was defined by aspiration of blood. The severity of the hemopericardium is also determined by ultrasound. Thus, when the echo-free space between the visceral and parietal pericardium was 10 to 20 mm, it was classified as moderate; and more than 20 mm was classified as severe pericardial effusion [8]. Nontraumatic hemopericardium was based on the condition that the patient with hemopericardium did not have any history of trauma in recent days according to the families’ statements or there were no signs of trauma in the field according to emergency medical technicians’ report. The procedures of SP or PCD were performed by board- certified emergency physicians in the ED. The method of Pericardial drainage was based on the personal preference of the treating physician as there was no randomization in this study. Effective decompression was defined as drainage of the pericardial blood of more than 100 mL in 5 minutes without difficulty, such as SP or PCD smoothly. In cases in which spontaneous circulation is restored, patients are transported to an operating room after consulting a cardiovascular surgeon, provided their families agree with the operation. Detailed information of patients in the final cohort was obtained retrospectively after patients discharge

through medical record abstraction by one of 2 physicians with extensive experience in chart review procedures. The following data were prospectively recorded for each patient: age, sex, comorbidities, bystander witness, relieving meth- ods of hemopericardium, effective decompression or not, initial rhythm on arrival, interval between calls to 911 and ambulance arrival on the ED, and outcome. Patient outcome was recorded as survival to hospital discharge and the ROSC, as primary endpoints. Effective decompression was recorded as a secondary end point.

Relieving method of hemopericardium

Percutaneous PCD

The PCD technique uses an 8-cm, 18-gauge needle. The needle is inserted between the xiphoid process and the left costal margin, and is aimed toward the left shoulder under ultrasound guidance. When the pericardial sac is entered, a guide wire is advanced through the catheter followed by an 8.5-F pericardial catheter.

Subxiphoid pericardiotomy

In the ED, we made a 5-cm-long skin incision from the lower end of the sternum caudally. The upper linea alba is divided in the midline with partial splitting of the xiphoid process while advancing to the pericardium by finger dissection. When touching the pericardium with an index finger, an inferior site of the pericardium is cut by scissors. A suction system was used for blood aspirate and clot evacuated. Then, a 32-F drainage tube is placed in the pericardial space.

Data analysis and statistics

We stratified patients into one of 4 groups with the relieving methods of hemopericardium. Group A, as direct SP group, classified as patients who initially underwent SP. Group B, as effective PCD group, classified as patients who initially underwent PCD smoothly. Group C, as delayed SP group, was classified as patients who initially underwent PCD with difficulty, and, subsequently, SP was performed. Group D, as ineffective decompression group, was classified as patients who initially underwent PCD with difficulty, but SP was not performed. The reasons why SP was not performed in Group D included refusal of the families and emergency physicians’ incompetence in SP. Appropriate sample size was calculated based on the assumption of a difference of 33% in ROSC between the groups. The difference was considered significant, and a sample size of 21 patients was needed to prove this difference (? set at .05; ? set at .2; power, 80%). All analyses were performed on SPSS 12.0 for Windows (SPSS, Inc, Chicago, Ill). Continuous variables were presented as median (25th, 75th percentile) and categorical variables as frequencies and percentages. Comparisons between the groups were made

with Pearson ?2 test for the categorical variables and Kruskal-Wallis test for continuous variables using the Mann-Whitney U test for post hoc analysis to adjust for multiple testing. We also used multiple logistic regression analysis adjusted by age, sex, and bystander witness to see whether the relieved method of hemopericardium was an independent predictor of ROSC in OHCA patients with nontraumatic hemopericardium. Statistical significance was set at P b .05 (2-tailed).

Results

Baseline population characteristics

During our study period, a total of 1491 OHCA resuscitation records were collected. Of these, 44 patients had documented pericardial effusion. The incidence of PEricardial effusion in OHCA patients who were transported for resuscitation is 3% (44/1491). Twenty-one patients were excluded from the study because of major trauma (n = 4), small amount of pericardial effusion (n = 7), did not perform confirmed pericardial diagnostic tapping (n = 2), history of malignancy (n = 1), previous documented pericardial effusion (n = 1), and signed DNAR order (n = 6). The final study cohort consisted of 23 patients with 15 (65.2 %) men and 8 (34.8 %) women. Their ages ranged from 47 to 94 years (mean, 72.0 years; median, 70 years). The cohort was divided into 4 groups according to the relieving methods of hemopericardium. Four (17.4 %) patients were placed in group A, 7 (30.4 %) in group B, 6 (26.1 %) in group C, and 6 (26.1 %) in group D. A flow diagram illustrating patient enrollment and group allocation is shown in Fig. 1. The general characteristics of the cohort are summarized in Table 1. There was no statistically significant difference in age, sex, comorbidities, bystander witness, initial rhythm on arrival, interval between calls to 911 and ambulance arrival on the ED between 4 groups (Table 1). Although neither comorbidities nor bystander witness had any statistical differences, they appeared to be clinically different among the groups. Pulseless electrical activity and asystole constituted the majority of the initial rhythms in these patients (22/23, 95.6%).

Return of spontaneous circulation rate and the rate of survival to hospital discharge

There was a clear difference in the ROSC rate between the 4 groups (P b .05). The direct SP group had the highest ROSC rate, 100% (4/4); the effective PCD group, 42.9% (3/7); the delayed SP group, 33.3% (2/6); and the ineffective decompression group, 0% (0/6). We also used ?2 test for trend to analyze the trend effect of early effective decompression. Patients who had early effective decom- pression tended to have higher ROSC rate (Fig. 2). The

Fig. 1 Patient enrollment and group allocation.

overall ROSC rate was 39.1% (9/23).The rate of survival to hospital discharge of the direct SP group, the effective PCD group, the delayed SP group, and the ineffective decom- pression group was 25% (1/4), 0% (0/7), 0% (0/6), and 0% (0/6), respectively. There was no significant difference in the rate of survival to hospital discharge between the 4 groups. The overall rate of survival to hospital discharge was 4.3% (1/23). One patient (patient 3) had survival to hospital discharge with cerebral performance category 1 (good cerebral performance) [9] and the morbidity 3 months after discharge was scored 0 (no symptoms) according to the Modified Rankin scale [10]. The cause of her hemoperi- cardium was rupture of the lateral wall of the left ventricle. She also had a long-term survival for 5 years. During our study period, the average rate of ROSC and that of survival to hospital discharge in all OHCA patients were 31.7% and 8.1%, respectively.

Independent predicator of ROSC

A multiple logistic regression analysis adjusted by age, sex, and bystander witness was constructed. Relieving methods of hemopericardium was an independent pre- dictor of ROSC in OHCA patients with nontraumatic hemopericardium (odds ratio, 0.17; 95% confidence

interval, 0.4-0.70).

Ineffective rate of decompression

In our study, there were 23 patients in which 29 procedures (10 SP and 19 PCD) were performed. The ineffective rate of decompression in SP and PCD was 0% (0/ 10) and 63.2% (12/19), respectively. There was a statistically significant difference in adequate relief of hemopericardium based on the 2 techniques applied (P b .01).

The ED subxiphoid pericardiotomy

The detailed characteristics of the patients who received SP are shown in Table 2. A total of 10 patients received SP. The median age was 70.5 years (range, 65.8-83 years). Seven (70%) patients were male. Most of them were witnessed arrest (8/10; 80%). Three patients who were brought to our ED by their families were in arrest at the time of arrival; thus there was no documented prehospital time interval (patients 3, 8, and 9). The average interval between calls to 911 and ambulance arrival on the ED for the 7 remaining patients was

21.3 minutes. The ROSC rate in patients who underwent SP was 60% (6/10). The rate of survival to hospital discharge in patients who underwent SP was 10% (1/10). Only 3 patients were sent to an operation room after ROSC (patients 2, 3, and 7). Three patients were not sent to the operation room after ROSC because their families refused (patients 1 and 8) and because of cardiovascular surgeon’s decision (patient 4). The hemopericardium mixed with blood and Blood clots in 8 patients (patients 1, 3, 4, 5, 6, 7, 9, and 10).

Discussions

To our knowledge, this is the first study that compares the relieving methods of hemopericardium, which focused on the nontraumatic OHCA patients with hemopericardium solely. The incidence of nontraumatic hemopericardium in OHCA patients is low in previous studies [6,11]. In a 10-year consecutive medicolegal autopsy by Altun et al [11], 40 (7%) of 569 deaths were related to hemopericardium. Of these, traumatic injury is the major cause of hemopericar- dium (38/40), whereas only 2 patients are nontraumatic. The incidence of nontraumatic hemopericardium in the study by Altun et al is 0.4% (2/569). In a 5-year study by Kurimoto et al [6], there were 76 nontraumatic OHCA or near cardiac arrest patients with hemopericardium, but they did not delineate the incidence. In our study, the incidence of pericardial effusion in OHCA patients was 3% (44/1491). The incidence of nontraumatic hemopericardium in OHCA patients ranged between 1.5% and 3% (23-44/1491). The difference in incidence of nontraumatic hemopericardium between our study and that by Altun et al may be due to differences between the study populations and differences in methodology.

Characteristic	Total (N = 23) Group A (n = 4) Group B (n = 7) Group C (n = 6) Group D (n = 6)			P a
Age, median (25th, 75th percentile), y	70.0 (65.0, 82.0) 75.5	(67.5, 85.0) 73.0 (48.0, 81.0) 70.5	(60.3, 82.0) 69.5 (68.0,85.0)	.86 b
Male sex, n (%)	15 (65.2) 2	(50) 4 (57.1) 5	(83.3) 4 (66.7)	.70
Comorbidities, n (%)
Diabetes	5 (21.7) 1	(25) 0 (0) 2	(33.3) 2 (33.3)	.54
Hypertension	13 (56.5) 1	(25) 5 (71.4) 3	(50) 4 (66.7)	.12
CAD	4 (17.4) 1	(25) 0 (0) 1	(16.7) 2 (33.3)	.35
CVA	2 (8.7) 0	(0) 1 (14.3) 0	(0) 1 (16.7)	.56
Bystander witness, n (%)	12 (52.2) 3	(75) 3 (42.9) 5	(83.3) 1 (16.7)	.09
Prehospital interval, c median	21.0 (17.5, 24.0) 22.0	(17.0, 26.0) 18.0 (12.0, 21.8) 22.0	(15.3, 25.8) 21.5 (17.5, 25.3)	.66 b
(25th, 75th percentile), min
Initial rhythm on arrival, n (%)
VF/pulseless VT	1 (4.3) 0	(0.0) 0 (0.0) 1	(5.3) 0 (0.0)	.43
PEA	10 (43.5) 3	(75.0) 3 (42.3) 3	(50.0) 1 (16.7)	.32
Asystole	12 (52.2) 1	(25.0) 4 (57.1) 2	(33.3) 5 (83.3)	.22
Massive hemopericardium, n (%)	18 (78.3) 4	(100.0) 5 (71.4) 4	(66.7) 5 (83.3)	.60
ROSC, n (%)	9 (39.1) 4	(100.0) 3 (42.9) 2	(33.3) 0 (0.0)	.02
Survival to hospital discharge, n (%)	1 (4.3) 1	(25.0) 0 (0.0) 0	(0.0) 0 (0.0)	.17
CAD indicates coronary artery disease; CVA, cerebrovascular accident; VF, ventricular fibrillation; VT, ventricular tachycardia. a Univariate comparisons between the 4 groups. b P values using the Mann-Whitney U test for post hoc analysis to adjust for multiple testing are not significant. c Interval between calls to 911 and ambulance arrival on the ED.

Previous studies have extensively investigated the managements of pericardial effusion. However, most of these studies had populations who were in-patients receiv- ing elective procedures with etiologies such as malignancy, uremia, connective tissue diseases, and iatrogenic cardiac perforation [2-5]. There are several drainage methods, for instance, needle pericardiocentesis, conventional subxi- phoid pericardiostomy, PCD, SP, sternotomy, and pericar- diectomy [6,12-14]. Blind pericardiocentesis or PCD has many complications. Laceration or perforation of the myocardium and the coronary vessels, pneumothorax, peritoneal cavity puncture, and trauma to internal mammary vessels, etc, all had been mentioned in the literature [15]. Echocardiographic guidance reduced the rate of complica- tions to 4.7% [4]. Ultrasound-guided procedure can be performed quickly in the ED [6]. Incidence of major complications is also reduced by using fluoroscopic guidance [16], but it seems not suitable for critical patients in an ED setting. Surgical decompression like sternotomy and pericardiectomy is relative time consuming and is better performed in the operating room. Allen et al [2], summarized a series of studies on patients with benign and Malignant pericardial effusion and concluded that the combined mortality and morbidity rate of subxiphoid pericardiostomy was 0.6% and 1.5%, and those of PCD was 1.9% and 10.6%, respectively, in the 20th century. Limitations of SP include the need for surgically trained physicians, use of general anesthesia, and time-consuming to perform. Pericardial catheter drainage is less invasive. It can be performed in hemodynamic unstable patients. Drawbacks of it are higher morbidity and mortality [2].

Table 1 General characteristics of 23 patients and univariate comparisons between the 4 groups

Moreover, it cannot guarantee effective drainage especially in the circumstances of hemopericardium where blood clot formation would be very quick and hazardous to patients with circulatory collapse [6].

For cardiac arrest patients, timing is critical. Subxiphoid pericardiotomy and PCD have been used in a minute for emergency cases in the ED and do not interrupt the chest compressions during resuscitation [6]. In our study, we used SP or PCD to relieve the hemopericardium. The average time in completing SP was within 3 minutes without specialized surgical instruments, and the chest compressions during the procedure were nearly not interrupted. The effective rate of decompression in SP was 100% (9/9). The average time in completing the PCD was also within 3 minutes, and the chest compressions during the procedures were nearly not interrupted either. But the effective rate of decompression in PCD was 36.8% (7/19). There was a statistically significant difference in adequate relief of hemopericardium based on the 2 techniques applied. Of these, although 6 patients were soon shifted to SP (group C), the ROSC rate was lower compared to that in the effective PCD group (group B) (33.3% vs 42.9%). Furthermore, the direct SP group (group A) has a higher ROSC rate than the effective PCD group (group B), probably because of the higher ability of SP to evacuate blood clots in the pericardial sac. From group A to D, it seems that delayed effective decompression lowered the ROSC rate. In our study, relieving methods of hemopericardium was an independent predictor of ROSC in OHCA patients with nontraumatic hemopericardium. Hemo- pericardium is distinct from other kinds of pericardial effusions; direct SP tends to have a higher ability to perform

Case No.

Age, y

Sex

Bystander witness

Prehospital interval, a min

initial cardiac rhythm

Severity of PE

PCD attempted

Cardiac rhythm after effective decompression

Blood aspirated, mL

Blood clots

ROSC

Operation

Prognosis

1

86

M

N

26

Asystole

Massive

N

Sinus rhythm

2800

Y

Y

N

Died

2

67

M

Y

22

PEA

Massive

N

Sinus rhythm

1250

N

Y

Y

Died

3

82

F

Y

–

PEA

Massive

N

Sinus rhythm

1000

Y

Y

Y

Survived

4

69

F

Y

17

PEA

Massive

N

Sinus rhythm

1500

Y

Y

N

Died

5

78

F

N

25

Asystole

Massive

Y

Asystole

480

Y

N

N

Died

6

72

M

Y

14

PEA

Massive

Y

PEA

180

Y

N

N

Died

7

55

M

Y

19

VF

Massive

Y

Sinus rhythm

1350

Y

Y

Y

Died

8

69

M

Y

–

PEA

Massive

Y

Sinus rhythm

2000

N

Y

N

Died

9

94

M

Y

–

PEA

Moderate

Y

PEA

500

Y

N

N

Died

10

62

M

Y

26

Asystole

Moderate

Y

PEA

330

Y

N

N

Died

F indicates female; M, male; N, no; Y, yes; PE, pericardial effusion. a Interval between calls to 911 and ambulance arrival on the ED.

early effective decompression. Likewise, Kurimoto et al [6] compared PCD and blind SP, and concluded that blind SP is safe and effective in ED patients with hemopericardium.

Table 2 Summary of initial presentation, management, and outcome of the 10 OHCA patients with hemopericardium who underwent SP

Along with effective decompression, volume replacement and noninterrupted chest compressions are also very important in resuscitation to OHCA patients with nontrau- matic hemopericardium. Usually, the evacuated blood was large in massive hemopericardium, given effective decom- pression. Tsang et al [5] reported that the Hemorrhagic fluid drainage volume ranges from 8 to 1700 mL, with an additional 0 to 830 mL in the following days. After effective decompression with SP and ensuing ROSC, the patients may tend to become hypovolemia owing to rebleeding. In our study, 6 patients who underwent SP had the blood aspirated to more than 1000 mL. We suggest patients have 2 or more large bore catheters or even trauma kit for fluid and blood transfusion. The definite treatment is surgical intervention. It could allow various kinds of repair of cardiac or great vessels injury. The whole resuscitation team should be very familiar with the core concepts in managing these patients: early advanced cardiac life support, early effective decompression, adequate fluid resuscitation or blood component transfusion, and early surgical intervention. There is the possibility that there will be only 1 survivor in our study, which may partly be due to having a not well-enough organized, linked team in the resuscitation. The long pre-hospital time is probably a factor. There still much effort to be done to rescue these critical patients.

The prognosis of pericardial effusion is quite variable and largely depends on the Underlying etiology as long as patients’ hemodynamic status is not life threatening [17]. Levine et al [18] found that a cumulative Probability of survival at 1 year for the group with malignancy and without is 17% and 91%, respectively. Furthermore, the prognostic variability is much larger in different groups with hemoper- icardium. For instance, hemorrhagic pericardial effusion has been correlated with poor outcome if it is related to

malignancy in the study by Colombo et al [19]. However, Atar et al [20] reported that iatrogenic hemopericardium generally has good prognosis; none died on a long-term follow-up of 33 +- 21 months [20]. Successful rescue pericardiocentesis for iatrogenic tamponade is 99% and early mortality within 1 month is less than 1% [5]. On the other hand, Isselbacher et al [21] identified acute aortic dissection patients with cardiac tamponade. The early mortality among those patients presenting with hypotension or normotension was 43% and that with cardiopulmonary arrest was 100%. Bayegan et al [7] found that patients with Acute type A aortic dissection and signs of preoperative cardiac tamponade without palpable pulses have an up to 16-fold higher risk of preoperative death, even when aggressive treatment with pericardiocentesis or surgical

Fig. 2 Rate of ROSC in the 4 groups. P b .05 (?² test for trend).

drainage is initiated. Hemopericardium in OHCA patients generally has poor prognosis. In the study of Kurimoto et al [6], the survival rate was 25% in the blind SP group and 9% in the PCD group in mixed cardiopulmonary arrest or near cardiopulmonary arrest population with hemopericardium. In our study, the overall rate of survival to hospital discharge was 4.3% (1/23). But the rate of survival to hospital discharge was 25% (1/4) in the direct SP group. During our study period, the average rate of survival to hospital discharge in all OHCA patients was 8.1%. It means that the prognosis in OHCA patients with nontraumatic hemo- pericardium is poor compared to those without hemoper- icardium; however, if early effective decompression was performed, their prognosis might be better.

This study has some limitations. First, this study is limited by the prospective-retrospective drawback typically occurring in resuscitation studies. Although analyzed retro- spectively, the data we prospectively collected were objectively detailed in the formal resuscitation record and were in accord with the Utstein recommendations [22]. This can minimize the recall bias and the possibility of missing data. Second, the number of patients in each group is small partly due to the low incidence of OHCA patient with nontraumatic hemopericardium. Third, we did not prospec- tively randomize the patients into different treatment groups owing to concern with ethical problems. Fourth, because not all patients received SP, the incidence of mixed blood and blood clots may be underestimated or overestimated. Fifth, the selection biases, such as comorbidities and witnessed arrests, appear to be clinically different among the groups and influenced both the approach and the treatment after ROSC. Moreover, as autopsy after death is not generally accepted in our culture, the exact cause of hemopericardium remains unknown.

Conclusions

Despite its inherent limitations, this study suggests that early and effective decompression is helpful in the OHCA patients with nontraumatic hemopericardium. Relieving methods of hemopericardium was an independent predictor of ROSC in OHCA patients with nontraumatic hemoper- icardium. In particular, SP has better effective decompression of hemopericardium than PCD. Further large, multi-center study is needed to confirm these results. We think the data afforded by this study may help improve the initial manage- ment of OHCA patients with nontraumatic hemopericardium.

Acknowledgments

We thank Professor Chi-Ren Hung, Dr Ming-Jen Lu, Dr Chia-Hsun Lin and Dr Hung-Hsing Chao for clinical backup and technical support.

References

1. Shabetai R. Pericardial effusion: haemodynamic spectrum. Heart 2004;90:255-6.
2. Allen KB, Faber LP, Warren WH, Shaar CJ. Pericardial effusion: subxiphoid pericardiostomy versus percutaneous catheter drainage. Ann Thorac Surg 1999;67:437-40.
3. Becit N, Unlu Y, Ceviz M, Kocoqullari CU, Kocak H, Gurlertop Y. Subxiphoid pericardiostomy in the management of pericardial effusions: case series analysis of 368 patients. Heart 2005;91: 785-90.
4. Tsang TS, Enriquez-Sarano M, Freeman WK, et al. Consecutive 1127 therapeutic echocardiographically guided pericardiocenteses: clinical profile, practice patterns, and outcomes spanning 21 years. Mayo Clin Proc 2002;77:429-36.
5. Tsang TS, Freeman WK, Barnes ME, Reeder GS, Packer DL, Seward JB. Rescue echocardiographically guided pericardiocent- esis for cardiac perforation complicating catheter-based proce- dures. The Mayo Clinic experience. J Am Coll Cardiol 1998;32: 1345-50.
6. Kurimoto Y, Hase M, Nara S, et al. Blind subxiphoid pericardiotomy for cardiac tamponade because of acute hemopericardium. J Trauma 2006;61:582-5.
7. Bayegan K, Domanovits H, Schillinger M, Ehrlich M, Sodeck G, Laggner AN. Acute type A aortic dissection: the prognostic impact of preoperative cardiac tamponade. Eur J Cardiothorac Surg 2001;20: 1194-8.
8. Weitzman LB, Tinker WP, Kronzon I, Cohen ML, Glassman E, Spencer FC. The incidence and natural history of pericardial effusion after cardiac surgery. An echocardiographic study. Circulation 1984;69:506-11.
9. Brain Resuscitation Clinical Trial I Study Group. A randomized clinical study of cardiopulmonary-cerebral resuscitation: design, methods, and patient characteristics. Am J Emerg Med 1986;4: 72-86.
10. van Swieten JC, Koudstaal PJ, Visser MC, Schouten HJ, van Gijn J. Interobserver agreement for the assessment of handicap in stroke patients. Stroke 1988;19:604-7.
11. Altun G, Altun A, Yilmaz. Hemopericardium-related fatalities: A 10-year medicolegal autopsy experience. Cardiology 2005;104:133-7.
12. Okamato H, Shinkae T, Yamakido M, Saijo N. Cardiac tamponade caused by primary Lung cancer and the management of pericardial effusion. Cancer 1993;71:93-8.
13. Press O, Livingston R. Management of pericardial effusion and tamponade. JAMA 1987;257:1088-92.
14. Miller JI, Mansour KA, Hatcher CR. Pericardiectomy: current indications, concepts and results in a university center. Ann Thorac Surg 1982;34:40-5.
15. Kennedy UM, Mahony NJ. Acadaveric study of complications associated with the subxiphoid and transthoracic approaches to emergency pericardiocentesis. Eur J Emerg Med 2006;13:254-9.
16. Maisch B, Ristic AD. Tangential approach to small pericardial effusions under fluoroscopic guidance in the lateral view: the halo phenomenon [abstract]. Circulation 2001;103(suppl A):II-730.
17. Soler-Soler J, Sagrista-Sauleda J, Permanyer-Miraldda G. Manage- ment of pericardial effusion. Heart 2001;86:235-40.
18. Levine MJ, Lorell BH, Divier DJ, Come PC. Implications of echocardiographically assisted diagnosis of Pericardial tamponade in contemporary Medical patients: detection before hemodynamic embarrassment. J Am Coll Cardiol 1991;17:59-65.
19. Colombo A, Olson HG, Egan J, Gardin JM. Etiology and prognostic implications of a large pericardial effusion in men. Clin Cardiol 1988; 11:389-94.
20. Atar S, Chiu J, Forrester JS, Siegel RJ. Bloody pericardial effusion in patients with cardiac tamponade. Is the cause cancerous, tuberculous, or iatrogenic in the 1990s? Chest 1999;116:1564-9.
21. Isselbacher EM, Cigarroa JE, Eagle KA. Cardiac tamponade complicating Proximal aortic dissection is pericardiocentesis harmful? Circulation 1994;90:2375-8.
22. Cummins RO, Chamberlain D, Hazinski MF, et al. Recommended guidelines for reviewing, reporting, and conducting research on in-

hospital resuscitation: the in-hospital ‘Utstein style‘. A statement for healthcare professionals from the American Heart Association, the European Resuscitation Council, the Heart and Stroke Foundation of Canada, the Australian Resuscitation Council, and the Resuscitation Councils of Southern Africa. Resuscitation 1997;34:151-83.