Background

The pathophysiology of cardiac tamponade (CT), as a cause of death, is related to an increase in intrapericardial fluid pressure that exceeds atrial venous pressures, thereby impeding venous return to the heart[1]. Rapidly evolving HP (200 to 300 ml) is more likely to cause death from CT than slowly evolving pericardial fluid accumulation (500 to 2000 ml), the latter allowing for accommodation of greater volumes due to gradual distension of the pericardial sac[1]. The normal volume of pericardial fluid (30 to 50 ml) reflects a balance between production and reabsorption[1].

The causes of CT include active or passive pericardial effusion and haemopericardium (HP) consequent on trauma, iatrogenic intervention or, either rupture of an acute myocardial infarction (RAMI) or intrapericardial rupture of a dissecting ascending aortic aneurysm (RD3A)[1, 2]. HP commonly follows RAMI, RD3A or trauma [1–3]; it has also been described in association with malignancy[3, 4], chemotherapy[5], homicide[6], pacing wire[7] or central venous catheter[8, 9] insertion, endocardial biopsy[10], open heart surgery[11], interventional coronary artery procedures (atheroablative, angioplastic or stenting)[12], exercise stress test[13], electroconvulsive therapy[14], coronary artery vasculitis[15] or dissecting aneurysm[16], myocardial abscess[17], infective endocarditis[3] and, during amniocentesis[18] or in the prenatal period[19]. CT been described in association with pneumopericardium [20] and as a complication of intrapericardial drain insertion[3]. HP has been described as a complication of 5% to 10% of patients with AMI [21][22].

Post-AMI myocardial rupture includes ventricular free wall rupture, ventricular septal rupture or papillary muscle rupture [1, 3]. Risk factors for this include age >60 yrs, female gender, pre-existing hypertension and lack of left ventricular wall hypertrophy[1].

The lateral wall, at mid-ventricular level, is said to be the most common site for post-infarction free-wall rupture[1]. However, the lateral and inferior aspects of the left ventricle have been reported as equally susceptible to post-infarctive rupture [21][23].

We have reviewed all cases of patients dying of cardiac tamponade in our hospital over a ten-year period to determine the causes of cardiac tamponade and highlighted some associated morphological features.

Methods

Our department performs postmortems at the request of Her Majesty (HM) Coroner, in the County of Cornwall, that has a population of ~430,000. Almost all deaths in this catchment area, falling within the jurisdiction of HM Coroner, are sent to our mortuary for a postmortem. These deaths are principally those within the community, or hospital, that fall within the purview of HM Coroner [24]: for the most part, these postmortems are performed because there is no firmly established cause of death.

The Mortuary Death Register of the Department Histopathology, Royal Cornwall Hospital, was reviewed for causes of death listed as either "haemopericardium" or "cardiac tamponade" between 1995 and 2004 (inclusive). These results were compared to a search of our textual postmortem report database using "moper" (a unique string of letters in the word "haemopericardium") and "tamponade" as separate search criteria within the text of the reports; the results of these were amalgamated and duplicates discarded.

All postmortem reports were retrieved and the cause of death confirmed as being attributable to HP or CT; the following parameters were then confirmed and retrieved from each postmortem:

1. The gender and age of patients

2. The causes of the CT or HP

3. If HP due to myocardial infarction, the

a. Presence of coronary artery thrombosis

b. Site of myocardial rupture and

4. Volume of intrapericardial haemorrhage

Results

14,368 postmortems were performed at the Royal Cornwall Hospital in the ten years under review. There were 461 cases of CT (3.2% of all postmortems), of which there were three cases of non-haemorrhagic effusion (passive = 2; infective = 1), five cases of post-traumatic HP and four miscellaneous causes (Table 1); these were excluded from subsequent analysis. Of the rest (n = 449; 98%), CT was due HP (Figure 1) secondary to RAMI (n = 311; 69%) or RD3A (n = 138; 31%) (Figure 2).

Table 1 Other causes of haemopericardium.
Full size table
Figure 1
figure 1

This image depicts a) the characteristic bluish black pericardial distension, observed at postmortem, most often b) containing an admixture of clotted and frank blood.

Full size image
Figure 2
figure 2

This image shows the a) intrapericardial site of rupture of a dissecting aneurysm of the ascending aorta, close to its origin and b) the presence of blood within such a dissection.

Full size image

Death due to HP, as a consequence of either RAMI or RD3A, is commoner in women than men (Figure 3). RAMI is a more common cause of HP than RD3A in both men and women.

Figure 3
figure 3

Cumulatively, more women die from RAMI or RD3A than men do.

Full size image

The age range of non-traumatic HP in men (33 to 99 years) and women (21 to 99 years) are similar. The mean age of men dying of HP/CT as a result of RAMI (73.4 ± 8.6) (n = 120) and RD3A (72.5 ± 13.0) (n = 62) is less than those of women dying of HP/CT from RAMI (78.7 ± 9.0) (n = 191) and RD3A (76.6 ± 11.1) (n = 76). The incidence of RAMI and RD3A in each age group is shown in Figures 4 &5; death from HP due to either cause is commoner in men than female before 70 yr but reverses after this and is commoner in women.

Figure 4
figure 4

Men die from RAMI more frequently, than women do, before the age 70 years than after it.

Full size image
Figure 5
figure 5

More women, than men, die from RD3A after the age of 70 years, than before it.

Full size image

RAMI

RAMI was responsible for most causes of death (69%; n = 311) due to HP, with an overall male (n = 120; 39%) to female ratio (n = 191; 61%) of 1:1.6.

Coronary artery thrombosis (CAT) was recognised in 210 postmortems (67%: 82 M, 128 F); CAT was not identified in the remainder (n = 101: 38 M, 63 F). There were no significant differences in the ages of men or women with CAT as a cause of RAMI.

The site of ventricular rupture was defined in all postmortem reports (table 2), with the anterior wall (33.8%) being the most common site of rupture. Cumulative figures show anterior (42.8%) and posterior ventricular wall rupture (37.3%) being more common than apical (6%) or lateral (22.5%) ventricular wall rupture. These figures are probably skewed by the larger numbers of females in the RAMI group.

Table 2 Site of rupture of acute myocardial infarction.
Full size table

Men were more likely to have posterior (12.2%) than anterior (10.6%) or lateral (7.4%) ventricular wall rupture; women, by contrast, had anterior (23.2%) ventricular wall rupture more commonly than posterior (19.3%) or lateral (5.8%) wall rupture.

RD3A, HP and CT

Patients dying from intrapericardial RD3A formed a smaller but significant group (30.5%; n = 138) of patients dying from HP. This mode of death was slightly more frequent in women (n = 76; 55%) than in men (n = 62; 45%) (Figure 3).

Intrapericardial blood volume

The volume of blood in the pericardial space was recorded in 246 (54%) instances (99 M, 147 F); this ranged between 150 ml and 1000 ml of blood in RAMI or RD3A (figure 6).

Figure 6
figure 6

This shows the volume of intrapericardial blood in each of 246 (index) cases where this was recorded.

Full size image

In addition, 800 ml of blood was seen in association with pericardial angiosarcoma, 750 ml in association with metastatic lung carcinoma and 1650 ml in association with malignant pericarditis. The volume of blood in association with Warfarin therapy was not recorded.

A difference in intrapericardial blood volume was noted between men and women whether as a consequence of either RAMI (M = 473 ml, F = 418 ml) or RD3A (M = 514 ml, F = 474 ml).

Death was associated with a slightly smaller volume of HP in-patients dying as a consequence of RAMI (mean = 440 ml) when compared to RD3A (mean = 498 ml).

Conclusion

There is little evidence in the literature reflecting on the causes or demographics of HP in routine postmortem practice. With the exception of individual case reports [4][6][7][9][10][11][12][13][14][15][16][17][18][19][20], most of the evidence available in standard textbooks is non-referenced.

Only 3.2% of all our deaths, over a 10-year period, were attributable to CT. In our series RAMI and RD3A were the principal causes of CT, causing death as a consequence of HP. Pericardial effusion, post-traumatic CT and miscellaneous causes of CT together formed only a small group of CT-related deaths.

Although the lateral wall at mid-ventricular level [1][21][23] or the inferior wall [21]are said to be the most common sites for post-infarction free-wall rupture [1][21][23], in our series anterior left ventricular wall rupture was more common than rupture at other ventricular wall sites. This might be a reflection of the greater numbers of women in our study, anterior free wall rupture being more common in women and posterior free wall rupture in men.

By contrast to traditional risk factors of female gender and age >60 years[1], men appear to die from CT due either RAMI or RD3A earlier than women. Indeed, death from either of these causes is commoner in men than women before, but not after, 70 years of age. This probably reflects the effects of cardiogenic risks in women evolving in the postmenopausal period and of men dying earlier from other causes, including AMI without myocardial rupture.

Traditionally, acute cardiac tamponade is associated with between 200 ml and 300 ml of sudden accumulation of intrapericardial fluid[1] or, in chronic slowly evolving accumulation, of volumes between 1000 and 2000 ml[1, 3]. This contrasts our experience with volumes of between 440 ml and 500 ml in association with CT due to acute pericardial accumulation of blood, in women and men respectively. In cases where HP volumes were larger than this, we speculate that this was associated with a slow leak, with or without a terminal phase of catastrophic rapid haemorrhage. We also demonstrated that slightly less intrapericardial blood volume is associated with RAMI than RD3A in fatal HP.

In conclusion we have defined the aetiology of CT in a postmortem population in rural South-West England and have described some of the salient morphological features associated with CT.