Pericardial Effusion Diagnosed by Echocardiography

Pericardial Effusion Diagnosed by Echocardiography

Pericardial Effusion Diagnosed by Echocardiography* Clinical and Electrocardiographic Findings in 171 Patients Marvin Berger, U.D., F.C.C.P.;** Leopol...

2MB Sizes 2 Downloads 71 Views

Pericardial Effusion Diagnosed by Echocardiography* Clinical and Electrocardiographic Findings in 171 Patients Marvin Berger, U.D., F.C.C.P.;** Leopold Bobak, M . D . ; f Mansoor Jelveh, M.D.;f and Emanuel Goldberg, M.D., F.C.C.P4

Clinical and electrocardiographic findings in 171 patients with pericardial effusion diagnosed by echocardiographic studies were reviewed. I n 70 patients the effusion was unsuspected. There were 87 small, 50 moderate, and 31 large effusions. Cardiac tamponade was present in three patients. Congestive heart failure was the most common cause of pericardial effusion and occurred in 37 patients. Other frequently noted conditions included cardiac disease without congestive heart furare, neoplasms, acute

nonspecific pericarditis, renal failure, and acute myocardial infarction. A pericardial friction rub was present in 23 patients, two-thirds of whom had moderate or large effusions. Atrial arrhythmias were common. L o w voltage occurred in 31 of 136 patients and was more common with large effusions. The ability to distinguish between a small effusion and the quantity of pericardial fluid present normally is a problem requiring further clarification.

p h e value of echocardiographic studies in the di••• agnosis of pericardial effusion is clearly established. Since 1965, when Feigenbaum1 first applied this technique to the recognition of pericardial effusion, numerous reports have appeared confirming echocardiographic studies as the safest, fastest, and most reliable method currently available for the detection of this condition.1*4 These reports have concerned themselves primarily with the technical aspects involved in using ultrasound to demonstrate the presence of pericardial effusion.5 In spite of the widespread application of echocardiographic studies, most of the information currently available regarding the etiology, incidence, and associated clinical findings in pericardial effusion is based on diagnostic methods that generally do not lend themselves to routine systematic investigation of large populations of patients. Most of these techniques (such as pericardial aspiration, angicardiographic studies, intravenous injection of carbon di-

oxide, and open thoracotomy), in addition to being time-consuming, produce discomfort for the patient and on occasion are associated with significant morbidity. 6 ( P P 2 9 8 - 3 0 4 ) Other techniques, such as cardiac fluoroscopic examination, are noninvasive but are probably less sensitive when dealing with small effusions and may present additional diagnostic difficulties in a patient with a dilated, poorly contracting left ventricle. Studies at autopsy are accurate but have limited application when attempting to establish the prevalence of pericardial effusions in various diseases. In the course of performing routine echocardiographic examinations in a large group of patients with a wide range of cardiac and systemic disorders, we have frequently noted the presence of pericardial effusions. Furthermore, these were often unsuspected clinically. These observations prompted us to review our experience with pericardial effusions diagnosed echocardiographically in an attempt to gather data that may have useful clinical application. By this study, we hope to further clarify ( 1 ) the prevalence of pericardial effusions during routine echocardiographic studies, (2) how often such effusions occur as an unexpected finding, (3) the association of a pericardial friction rub and its relationship to the size of the effusion, and (4) the nature and specificity of associated electrocardiographic findings.

r

*From the Division of Cardiology, Department of Medicine, Beth Israel Medical Center, New York, and the Department of Medicine, Mount Sinai School of Medicine of the City University of New York. ** Attending Physician, Department of Medicine, and Assistant Professor of Clinical Medicine. tFellow in Cardiology. jChief, Division of Cardiology, and Associate Professor of Clinical Medicine. Manuscript received November 10; revision accepted March 28. Reprint requests: Dr. Berger, 10 Nathan D. Perlman Place, New York City 10003

174

BERGER ET AL

CHEST, 74: 2, AUGUST, 1978

MATERIALS AND METHODS

6

Echocardiographic examinations were performed on 2,106 patients at Beth Israel Medical Center, New York, from May 1, 1974 through Dec 31, 1976. Pericardial effusion was diagnosed i n 178 patients. Of this group, 171 patients ( 154 inpatients and 17 outpatients) had records considered ade­ quate for analysis and constitute the population of patients for the present study. There were 80 male and 91 female patients ranging i n age from 11 to 87 years (mean, 57 years ) . Echocardiograms were obtained with an ultrasonoscope (Unirad series 100) using a 2.25-MHz 13-mm transducer focused at 7.5 cm with a repetition rate of 1,000 impulses per second. Recordings were made on a strip-chart recorder (Tektronix 174) at paper speeds of 25 or 50 mm/sec. The diagnosis of pericardial effusion was based on criteria de­ scribed by Horowitz et a l . A single band of moving pericar­ dial echoes recorded at low sensitivity with no separation between epicardium and pericardium was considered normal (no pericardial effusion). A pericardial effusion was believed to be present when an echo-free space was clearly visualized between the epicardium and pericardium and was associated with flattening of the pericardial echo relative to the epicardial echo. An attempt was made to estimate the size of the effusion using previously published guidelines. ' * A pericardial ef­ fusion was considered small if the separation between epi­ cardium and pericardium was limited to systole and early diastole and was associated with a flat pericardial echo or i f the separation extended throughout the cardiac cycle with a minimal separation during disastole ( Fig 1A ). There were 87 patients in this category. I f the separation was limited to systole or i f there was no pericardial flattening ( pattern C i of 5

5 7

8

Horowitz et a l ) , the echocardiogram was considered not indicative of pericardial effusion. When the epicardium was clearly separated from the pericardium, generally by at least 1 cm, the effusion was designated as moderate in size ( F i g I B ) . Of the 50 patients in this group, 23 also had a separation between the anterior wall of the heart and the anterior wall of the chest The effusion was considered to be large when at least a 2-cm separation existed between the posterior left ventricular wall and the pericardium and was associated with an anterior space ( F i g 1C). Thirty-one patients were consid­ ered to have large effusions. A n anterior echo-free space i n the absence of a significant posterior echo-free space was considered not indicative of pericardial effusion. I n an addi­ tional three patients, cardiac tamponade was diagnosed on the basis of clinical findings. I t should be noted that the amount of pericardial fluid is often difficult to measure echocardiographically, and the classification used here should be considered no more than a gross estimation of the size of the effusion. I n each case an attempt was made to determine whether or not the effusion was suspected prior to the echocardiogram from a careful review of the hospital chart. I f this point could not be clearly resolved from the record, the referring physi­ cian was contacted. The etiology of the effusion was carefully sought in each patient, based on information obtained from the clinical record, from the attending physician, and, i n some cases, at surgery or autopsy. On the basis of this information, each patient was placed into one of the cate­ gories of recognized diagnoses known to produce pericardial effusion. The effusion was attributed to congestive heart failure i f it occurred along with cardiomegaly, pulmonary rales, or pulmonary congestion on the chest x-ray film and a third heart sound i n the absence of one of the recognized causes of pericardial effusion. A n effusion occurring i n the 9

FIGURE 1. A (left). Separation between epicardium (EP) and pericardium (P) due to small pericardial effusion. Β (center), More prominent separation between epicardium and peri­ cardium, due to moderate pericardial effusion. C (right)', Marked separation between epicardium and pericardium i n association with substantial anterior echo-free space. These findings indi­ cate large pericardial effusion. E N , Endocardium; PE, posterior effusion: and A E , anterior effusion.

CHEST, 74: 2, AUGUST, 1978

PERICARDIAL EFFUSION DIAGNOSED DY ECHOCARDIOGRAPHY 175

clinical setting of an acute myocardial infarction was placed in one of two categories. When acute pericarditis with effusion appeared within the first four days of the acute episode, i t was attributed to the myocardial infarction. Episodes occurring beyond the second week after the acute episode were considered as manifestations of the postmyocardial infarction syndrome. The presence or absence of a pericardial friction rub was sought in every patient from a review of the hospital record. I n each case the rub was heard by at least two observers. A n attempt was made to determine whether there was any relationship between the presence of pericardial rub and the size of the effusion. Electrocardiograms recorded within 48 hours of the echo­ cardiogram were available in 136 patients. The presence of atrial fibrillation, atrial flutter, or atrial tachycardia was noted. This information, which was based on a single 12-lead ECG, should not be considered representative of the true prevalance of these arrhythmias in pericarditis. The maxi­ mum deflection of the QRS complex in the leads on the extremities and the precordial leads was measured. Low voltage was diagnosed i n accordance with criteria of the New York Heart Association and was regarded as low 10

when the largest deflection of the QRS complex on one side of the reference level ( P-R segment ) was 0.5 mv or less in all of the extremity leads and 1.5 mv or less in all of the pre­ cordial leads. Abnormalities of the S-T segment and Τ wave that were consistent with pericarditis were diagnosed using guidelines described by Spodick. 9

11

12

T a b l e 1—Etiologic

Cause of Effusion

RESULTS

Etiology

The various causes of pericardial effusion in the 171 patients are summarized in Table 1. Congestive heart failure was found most often and was present in 37 patients (22 percent). The majority of these cases were due to ischemic heart disease or valvular heart disease. In these 37 patients, the various types of cardiac disorders and the number of patients with each were as follows: ischemic, 11; valvular, 10; cardiomyopathy, 6; hypersensitive, 4; cor pulmonale, 2; infective endocarditis, 2; idiopathic hypertrophic

Classification of Pericardial Effusions in 171 Patients

No. (Percent)

Effusion Suspected Yes

A

Size of Pericardial -Effusion

No

Small

Moderate

ALarge

Tamponade

Congestive heart failure

37 (22)

22

15

18

15

4

Cardiac disease (no congestive heart failure)

30 (18)

5

25

22

6

2

Neoplastic disease

18 (11)

14

4

7

4

5

2

Acute nonspecific pericarditis

15 (9)

15

5

5

5

...

Renal failure

11 (6)

10

1

4

2

4

1

Acute myocardial infarction

11(6)

8

3

7

3

1

Infective endocarditis

5(3)

3

2

1

2

2

Connective tissue and hyper­ sensitivity disease

6(4)

2

4

3

1

2

Hypothyroidism

5(3)

4

1

2

1

2

Postmyocardial infarction syndrome

5(3)

5

1

2

2

Pregnancy

5(3)

No cardiac or systemic disease

5(3)

1

Trauma

4(2)

4

Pulmonary emboli

3(2)

1

2

2

Miscellaneous underlying systemic disease

3(2)

...

3

3

Tuberculosis

2(1)

2

1

1

Drugs (procainamide)

2(1)

2

1

1

Primary pulmonary hypertension

2(1)

1

1

1

Radiation

1 (0.6)

1

1

Chylous

1 (0.6)

1

1

Total

171 (100)

ιοί

176 BERGER ET AL

5

5

4

2

... 2

1

70

87



3 2

...

1

50

...

...

...

... ...

...

...

31

3

CHEST, 74: 2, AUGUST, 1978

subaortic stenosis, 1; and unknown, 1. The second most common diagnostic category in­ cluded 30 patients (18 percent) with a wide range of cardiac disorders but no clinical findings of con­ gestive heart failure and no evidence of any other underlying condition known to produce pericardial effusion. Most of these patients had valvular heart disease (13) or ischemic heart disease (9). The re­ maining seven patients were evenly distributed among cardiomyopathy, hypertensive heart disease, and congenital heart disease. Other relatively com­ mon causes of pericardial effusion included the fol­ lowing: neoplastic disease, 18 patients; acute non­ specific pericarditis, 15 patients; renal failure, 11 patients; and acute myocardial infarction, 11 pa­ tients.

suspected effusions. In this group of 30 patients, only five effusions (17 percent) were suspected prior to the echocardiogram. In contrast to this, when con­ gestive heart failure was present, the effusion was much more likely to be recognized, being suspected in 22 ( 59 percent) of 37 cases. Conditions known to be specifically associated with pericardial involve­ ment ( such as renal failure, neoplastic disease, acute nonspecific pericarditis, and myocardial infarction) tended to have the highest incidences of suspected effusions prior to the echocardiogram. Table 1 sum­ marizes the frequency of suspected vs unsuspected effusions in each of the various etiologic categories.

Of particular interest was a group of five female patients in the third trimester of pregnancy who were referred for routine echocardiographic study because of the presence of a systolic murmur. I n each of these cases, the murmur was believed to be innocent in nature, and in none of them was the effusion suspected prior to the echocardiogram. Five patients had no evidence of any cardiac or systemic disease.

A pericardial friction rub was heard in 23 patients. Of this group, ten patients had large effusions, six had moderate effusions, and seven had small effu­ sions. There was no relationship between the size of the effusion and the presence of a rub. Sixteen (70 percent) of the 23 patients had moderate or large effusions, indicating that even in the presence of a substantial quantity of pericardial fluid, a rub may be heard.

Incidence

of Unsuspected

Effusions

In 70 patients (41 percent), the pericardial effu­ sion was unsuspected prior to echocardiographic studies. Small effusions were much less likely to be recognized than moderate or large effusions (Table 2). Only 43 percent (37) of the 87 small effusions were suspected. This incidence of suspected effu­ sions increased to 70 percent (35/50) with moderate effusions and to 84 percent (26/31) in the group with large effusions. I n all three patients with car­ diac tamponade, the condition was identified clin­ ically prior to the echocardiogram. The group of patients with underlying organic heart disease but no congestive heart failure and none of the conditions known to produce pericardial effusion had a surprisingly large number of unTable 2—Relationship of Size to Incidence of Unsuspected Pericardial Effusions Size of

Total

Effusion

No.

Suspected

Unsuspected

Small

87

37

50

Moderate

50

35

15

Large

31

26

5

3

3

0

171

101

70

Tamponade Total

CHEST, 74: 2, AUGUST, 1978

Pericardial

Friction

Electrocardiographic

Rub

Findings

Sinus rhythm was present in 103 of the 136 pa­ tients with available ECGs. One patient had a ven­ tricular rhythm from a right ventricular endocar­ dial pacemaker. The remaining 32 patients had atrial arrhythmias, an incidence of 24 percent. These consisted of 30 patients with atrial flutter or fibrilla­ tion and two patients with atrial tachycardia. Sig­ nificant underlying cardiac disease was present in 29 of the 32 patients with atrial arrhythmias. Low voltage was noted in 31 (23 percent) of the 136 patients and was almost twice as common with large effusions as with small or moderate effusions. Low voltage was present in 14 of 67 patients with small effusions, in eight of 44 patients with moderate effusions, and in nine of 25 patients with large effu­ sions. Only eight patients demonstrated characteristic abnormalities of the S-T segment and Τ wave that were clearly typical of pericarditis. This group con­ sisted of five patients with acute nonspecific pericar­ ditis, two with renal failure, and one patient with a tachycardia-bradycardia syndrome and a large peri­ cardial effusion of undetermined etiology. Electrical alternans was seen in one patient, a 42-year-old woman with cardiac tamponade due to metastatic carcinoma of the breast.

PERICARDIAL EFFUSION DIAGNOSED BY ECHOCARDIOGRAPHY 177

DISCUSSION

Echocardiography has produced a major impact on the ability to detect the presence of a pericardial effusion. The technique is rapid, safe, and extremely sensitive, even in the presence of relatively small quantities of pericardial fluid. Despite the easy recognition of pericardial effusion, its incidence is not well defined Studies performed on adult human cadavers shortly after death have demonstrated that 20 to 60 ml (average, 20 to 25 ml) of pericardial fluid are usually present. Using careful technique, and then verifying the findings at surgery, Horowitz et al 5 found that as little as 16 ml of fluid could be identified by echocardiographic studies. Therefore, the detection of a very small effusion in a given patient may not necessarily be clinically significant and may merely represent fluid that is present under normal conditions. Although this may occasionally be the case, there is some evidence suggesting that pericardial effusions occur more frequently than is generally appreciated. I n 3,000 consecutive autopsies performed at the University of Minnesota, 226 or 7.5 percent of the cases were noted to have 50 ml or more of serous fluid in the pericardium. 18 I n the present study, a review of 2,106 echocardiograms revealed a pericardial effusion in 178 patients, an incidence of 8.5 percent Riba and Morganroth 14 found a pericardial effusion in 15 percent of routine echocardiograms. On the basis of these observations, it appears that pericardial effusions may be present in a substantial number of patients. Furthermore, significant numbers! of effusions are clinically unsuspected, 41 percent (70/171) in the present study and 61 percent in the group of patients reported by Riba and Morganroth. 14 There is a lack of agreement in the literature regarding the incidence and the importance of pericardial effusions in patients with congestive heart failure. According to Wood, 1 5 fluid in the pericardial cavity is infrequent and is rarely of clinical significance in cardiac failure. Spodick 6 ^ 2 8 8 ) stated that pericardial effusion is uncommon except in the presence of rheumatic heart disease. I n a study of autopsies performed on 50 patients with congestive heart failure, Paul et a l i e found that effusions were uncommon. Of the eight patients with greater than 50 ml of pericardial fluid, five had less than 100 ml of fluid, and no patient was found to have more than 250 ml of fluid. On the other hand, Winters et al, 1 7 using intravenous injection of carbon dioxide to detect fluid in the pericardial cavity, found a pericardial effusion in one-third of the patients with congestive heart failure. Rheumatic heart disease was the most frequent underlying cardiac 178

BERGER ET AL

disorder in this group. Stewart et al 1 8 studied 25 patients with congestive heart failure and cardiomegaly using angiocardiographic techniques. A pericardial effusion was demonstrated in 17 subjects and decreased in size in most of the patients after therapy, suggesting that cardiac failure was responsible for the development of the effusion. In the present study, congestive heart failure was the clinical condition most commonly associated with pericardial effusion, and such failure accounted for 22 percent (37/171) of the effusions diagnosed echocardiographically. More than half of these were moderate or large. Furthermore, effusions were seen with almost equal frequency in all types of heart disease, regardless of etiology. These findings have led us to the conclusion that congestive heart failure may produce a pericardial effusion in a substantial number of patients. Although the majority of these effusions do not appear to be hemodynamically significant, further clarification is needed in this area. The demonstration of a pericardial effusion in 30 patients with organic heart disease but no congestive heart failure was a somewhat unexpected finding. Winters and Soloff,19 using venous angiocardiographic studies, found effusions to be extremely unusual in cardiac patients when congestive heart failure was absent; however, when such effusions did occur, they were limited to those patients with rheumatic heart disease. These investigators19 attributed this observation to rheumatic involvement of the pericardium, although none of the patients had clinical evidence of active rheumatic disease at the time of the study. In the present report, most of the cardiac patients with pericardial effusion and no congestive heart failure had valvular heart disease, primarily of rheumatic origin; however, there was also a significant number with ischemic heart disease, hypertensive heart disease, and cardiomyopathy. Although the mechanism leading to the development of a pericardial effusion in these patients is obscure, the apparent high incidence in comparison to previous reports is due in part to the ability of the echocardiogram to detect small amounts of fluid that ordinarily would go unrecognized by previously used diagnostic techniques. The finding of a pericardial effusion in five female patients who were in the third trimester of pregnancy merits further comment I t is likely that this is merely a manifestation of the generalized retention of fluid that occurs during pregnancy. Unfortunately, none of these individuals underwent repeat echocardiograms after delivery to determine if the effusion had resolved. It is believed by some that a pericardial friction CHEST, 74: 2, AUGUST, 1978

rub is not heard in the presence of an effusion. This is contradictory to our findings. Two-thirds of the patients with a pericardial rub were found to have large or moderate effusions. Spodick20 attributes the persistence of a rub in the face of an effusion to intrapericardial adhesions or to the existence of areas of pericardium that remain uncovered by fluid. It is frequently stated that pericardial disease is a common cause of atrial arrhythmias. I n a clinicopathologic study of 144 patients with disturbance of cardiac rhythm, James21 found involvement of the sinus node in all 38 patients with pericarditis. Twenty-six of these had documented atrial ar­ rhythmias. This was attributed to the proximity of the sinus node to the epicardium. In contrast to these findings, Spodick,22 in a prospective study of 100 patients with acute pericarditis, found atrial arrhythmias in only seven patients. In all of these, significant underlying heart disease was present. Spodick22 concluded that pericarditis occurring in the absence of disease of the myocardium, cardiac valves, or coronary arteries was unlikely to produce atrial arrhythmias. I n the present study, major un­ derlying cardiac disease was present in 29 of the 32 patients in whom atrial arrhythmias were recorded. In the remaining three patients (rheumatoid peri­ carditis, acute renal failure, and metastatic carci­ noma), pericarditis may have been responsible for the arrhythmia. Unfortunately, the true incidence of atrial arrhythmias in pericardial effusion and their relationship to underlying cardiac disease could not be ascertained because the information was based on a single 12-lead ECG. We have recorded our observations in a large group of patients with pericardial effusion. The use of echocardiographic studies has demonstrated the relatively frequent occurrence of this condition. I n view of this, further studies are needed to develop guidelines that wilt be of use in determining the significance of pericardial effusion in a given clinical situation. Furthermore, many effusions that are de­ tected by echocardiographic studies are unsuspected clinically. Although many of these are small and require no further evaluation, a significant number are substantial in size. The ability to differentiate between a small effusion and the quantity of peri­ cardial fluid that is normally present is an additional problem requiring further clarification.

CHEST, 74: 2, AUGUST, 1978

A C K N O W L E D G M E N T : We would like to thank Mrs. Linda Keener for her valuable technical assistance. REFERENCES 1 Feigenbaum H , Waldhausen JA, Hyde L P : Ultrasound diagnosis of pericardial effusion. JAMA 191:711-714, 1965 2 Feigenbaum H , Zaky A, Waldhausen JA: Use of reflected ultrasound in detecting pericardial effusion. Am J Cardiol 19:84-90, 1967 3 Goldberg BB: Ultrasonic determination of pericardial effusion. JAMA 202:103-106,1967 4 Klein JJ, Segal B L : Pericardial effusion diagnosed by reflected ultrasound. Am J Cardiol 22:57-64,1968 5 Horowitz MS, Schultz CS, Stinson EB, et al: Sensitivity and specificity of echocardiographic diagnosis of pericar­ dial effusion. Circulation 50:239-247, 1974 6 Spodick D H : Chronic and Constrictive Pericarditis. New York, Grune and Stratton, Ine, 1964 7 Feigenbaum H : Echocardiography (2nd ed). Philadel­ phia, Lea and Febiger, 1976, pp 430-432 8 Teicholz L E : Echocardiographic evaluation of pericardial effusion. In Gramiak R, Waag RC (eds): Cardiac Ultra­ sound. St. Louis, CV Mosby Co, 1975, pp 109-110 9 Spodick D H : Differential diagnosis of acute pericarditis. Prog Cardiovasc Dis 14:192-209, 1971 10 Likoff W : Pericarditis complicating myocardial infarction. Am J Cardiol 7:69-72,1961 11 Dressler W : The post-myocardial infarction syndrome. Arch Intern Med 103:24-28, 1959 12 Criteria Committee of the New York Heart Association: Diseases of the Heart and Blood Vessels: Nomenclature and Criteria for Diagnosis. Boston, l i t t l e , Brown and Co, 1964, ρ 438 13 Wolfe L I : A n autopsy study of disease of die pericar­ dium. Permanente Foundation Med Bull 8:122-139, 1950 14 Riba A L , Morganroth J: Unsuspected substantial pericar­ dial effusions detected by echocardiography. JAMA 236: 2623-2626, 1976 15 Wood P: Disease of the Heart and Circulation (2nd ed). Philadelphia, JB Lippincott Co, 1956, ρ 679 16 Paul O, Vawter GF, Schweitzer A W , et al: Pathological changes in congestive heart failure. Arch Pathol 64:363¬ 381, 1957 17 Winters W L Jr, Carter B L , Stauffer H M , et al: Pericar­ dial effusion i n congestive heart failure. Dis Chest 40:82¬ 86, 1961 18 Stewart D , Carson PH, Bahler RC, et al: Presence of pericardial effusions i n heart failure. Circulation 35 (suppl2):243-244, 1967 19 Winters W L , Soloff L A : Pericardial effusion i n clinically inactive compensated rheumatic heart disease. Am J Med Sci 242:173-176, 1961 20 Spodick D : Acute Pericarditis. New York, Grune and Stratton, Ine, 1959, ρ 60 21 James T N : Pericarditis and the sinus node. Arch Intern Med 110:305-311, 1962 22 Spodick D H : Arrhythmias during acute pericarditis. A prospective study of 100 consecutive cases. JAMA 235: 39-41,1976

PERICARDIAL EFFUSION DIAGNOSED BY ECHOCARDIOGRAPHY 179