Constrictive Pericarditis Treated by Surgery

Abstract

We reviewed the records of 45 patients (mean age, 46.6 ± 14.9 yr; range, 21–84 yr) with a diagnosis of constrictive pericarditis who had undergone pericardiectomy from 1994 through 2006. Preoperatively, 2 of the patients (4.4%) were in New York Heart Association (NYHA) functional class I, 20 (44.4%) in class II, 22 (48.9%) in class III, and 1 (2.2%) in class IV. Pericardial calcification was detected in 20% of plain chest radiographs. Constrictive pericarditis was caused by tuberculosis in 22.2%, chronic renal failure in 8.9%, a history of sternotomy in 4.4%, and malignancy in 4.4%. The cause was idiopathic in 60% of the patients. Low-output state was the most common postoperative problem (22.2%).

The mean follow-up period was 40 ± 18 months (range, 3–144 mo). Three months postoperatively, only 1 of 43 available patients (2.3%) was in NYHA class III, while the rest were in class I (36 patients; 83.7%) or II (6 patients; 14%). The overall mortality rate was 4.4%: 1 patient with tuberculosis died of respiratory insufficiency while hospitalized, and 1 died of metastatic adenocarcinoma during follow-up.

Our results show that pericardiectomy remains an effective procedure in the treatment of constrictive pericarditis. Tuberculosis is still an important cause of constrictive pericarditis in Iran, despite intensive vaccination and use of antitubercular drugs.

Constrictive pericarditis is an uncommon cause of heart failure. Approximately 15% of patients with acute pericarditis experience cardiac tamponade. Effusive pericarditis is found in 5% of cases of acute tamponade, but only 1.2% of these patients develop chronic constrictive pericarditis. Tuberculosis (TB) is the cause of nearly 4% of acute pericarditis cases and 7% of cardiac tamponade cases. According to reports during the past decade, the incidence of purulent and TB pericarditis was 55/1,184 patients (4.6%).¹

In the past, idiopathic inflammation and TB constituted the most common causes of constrictive pericarditis. However, during the last 20 years there has been a shift in the causes of constrictive pericarditis: radiation and previous cardiac surgery currently are the prime suspects in constrictive pericarditis cases in the Western world.²

This retrospective study was conducted to determine the causes of constrictive pericarditis in our institution's patient population over the past 12 years, with a view in particular toward ascertaining whether TB would emerge as a significant cause.

Patients and Methods

All patients who had undergone pericardiectomy at Rajaee Heart Center from 1994 through 2006 were identified in our surgical database. Clinical and operative details were retrieved from the hospital notes. Patient follow-up was performed by the cardiac surgeons and cardiologists in the hospital clinic. Perioperative death was defined as death within 30 days of the operation or during the same hospital admission.

There were 45 patients, including 28 men (62.2%), with a mean age of 46.6 ± 14.9 years (range, 21–84 yr). The diagnosis of constrictive pericarditis was confirmed by clinical presentation, echocardiographic study, cardiac catheterization, and chest computed tomographic (CT) scan, as needed. During the procedure, tissue and fluid sampling—for aerobic and anaerobic cultures and histopathologic studies—was done for an evaluation of the causal factors. The diagnosis of TB was confirmed on the basis of clinical findings in combination with histopathologic features, including the presence of acid-fast bacilli in Ziel-Nelson tissue staining, typical granuloma and caseous necrosis, and bacteriologic studies using the polymerase chain reaction (PCR) test on the pericardial fluid or tissue for evidence of mycobacterium tuberculosis. Six patients were admitted with a diagnosis of TB, 4 had chronic renal failure, 2 had a history of open cardiac surgery, and 1 had a history of exposure to chemical weapons. The preoperative characteristics of the patients are shown in Table I.

TABLE I. Preoperative Characteristics of the 45 Patients

Dyspnea with or without peripheral edema or ascites was the most common symptom. The rates of pericardial rub, high jugular venous pressure, and pulsus paradoxus in physical examination were statistically insignificant. Eight patients had normal physical findings. Nineteen patients had atrial fibrillation at the time of surgery. Pericardial calcification on plain chest radiography was seen in 9 patients, and 8 had no unusual changes. The diaphragmatic, apical, and left borders of the cardiac shadow on the chest radiograph were the most common sites of calcification. Most of the radiographs showed areas spared from calcification; however, in 2 of the 9 patients, global pericardial calcification was seen on the radiograph. Chest CT scanning was performed in only 26 (57.8%) of the patients, but the presence of various degrees of pericardial calcification was patently high (19/26 patients; 73%). Pericardial thickening was reported in almost all of the patients, and 3 patients had more than 10 mm of thickness upon chest CT scanning.

The surviving patients were monitored from 3 to 144 months (median, 41 mo), and 3- and 6-month follow-up periods were completed for 43/44 (97.7%) and 38/44 (86.4%) of the patients, respectively. The preoperative hemodynamic and echocardiographic characteristics of the patients are presented in Tables II and III, respectively.

TABLE II. Preoperative Hemodynamic Characteristics in the 45 Patients

TABLE III. Preoperative Findings of the Echocardiographic Studies in the 45 Patients

Surgical Technique

The primary surgical goal was total pericardiectomy, including the resection of the anterior pericardium between the 2 phrenic nerves, the basal aspect of the pericardium over the diaphragm, the posterior part of the pericardium lying on the left and right ventricles, and the pericardium over the great arteries and both atria. In the patients who had thick, dense, and constrictive pericardium over the pulmonary veins, pericardial resection was also performed beyond the phrenic nerves. In these patients, an effort was made to save the phrenic nerves as pediculated tissue. The left phrenic nerve was sacrificed in 2 patients because of severe calcification and adhesion of the nerve to the pericardium. Moreover, attempts were made to decorticate the constrictive white, fibrotic, and thickened layer of the epicardium over the ventricles. Some patients could undergo only partial pericardiectomy as a consequence of inadequate exposure (in 2 patients approached via thoracotomy), high risk of coronary artery or myocardial damage, or severe bleeding. In such cases, the pericardium over the right atrium or superior and inferior venae cavae was left intact. In 2 patients, the pericardium could be resected only in patches; therefore, some islands of epicardium and pericardium were left intact.

Three patients required concomitant valve replacement. In these patients, the anterior wall of the pericardium was resected before the remaining pericardium and epicardium were resected under cardiopulmonary bypass (CPB). Except for the 2 patients who underwent partial pericardiectomy via left thoracotomy at the surgeon's discretion, all the others were approached via median sternotomy. The primary intention was pericardiectomy without CPB; nonetheless, the concomitant cardiac procedure (3 patients), hemodynamic instability during manipulation, severe surgical bleeding, and the need for sufficient exposure necessitated the institution of CPB.

Statistical Analysis

All of the continuous variables are expressed as mean ± SD and the categorical variables as percentages. The χ² and Student t tests were performed as appropriate. The Wilcoxon signed rank test was used to compare the New York Heart Association (NYHA) functional classes of patients preoperatively and postoperatively. A P value <0.05 was considered statistically significant.

Results

The overall mortality rate in this series was 4.4% (2/45). The only early death (2.2%) was that of a 38-year-old man, who had presented in NYHA class III and had complete destruction of his lungs due to pulmonary TB; he died on the 24th postoperative day because of pulmonary insufficiency, despite the fact that his phrenic nerves had been saved during the surgical procedure.

The late mortality rate was 2.3% (1/44) in our study. The man who died had presented in NYHA class II, and histopathology showed that the pericardium was involved with a metastatic adenocarcinoma; he died 7 months after surgery as a result of hepatic insufficiency due to metastatic adenocarcinoma of the colon.

The surgical approach included sternotomy in 43 patients (95.6%) and thoracotomy in the remaining 2 (4.4%). Cardiopulmonary bypass was performed in only 9 patients (20%). Severe hemodynamic instability in 4 patients (8.9%), concomitant valve replacement in 3 (6.7%), and severe iatrogenic intraoperative bleeding in 2 (4.4%) were the primary reasons for CPB use. In addition to pericardiectomy, 1 patient underwent aortic valve replacement; 1, mitral valve replacement; and 1, tricuspid valve replacement. Twenty-eight patients (62.2%) required resection of the fibrotic dense epicardium. The use of inotropic agents is shown in Figure 1.

Fig. 1 Comparison of the use of inotropic agents during the perioperative and late postoperative periods.

It is of interest that 22 patients (48.9%) had constrictive pericarditis associated with pericardial effusion. The most common early postoperative complication was low-output syndrome (10 patients). The mean volume of postoperative mediastinal bleeding was 433 ± 215 mL (range, 100–1,300 mL). One patient was re-explored due to excessive mediastinal hemorrhaging. The mean number of postoperative blood transfusions was 1.8 ± 0.7 units (range, 0–6 units). Table IV shows the early post-pericardiectomy complications.

TABLE IV. Early Postoperative Complications

The functional status of the patients improved after pericardiectomy. Forty-two of the 45 patients (93.3%)had been in NYHA class II or III before the operation, but 3 months postoperatively only 1 of 43 available patients (2.3%) was in NYHA class III, and the rest were in class I (83.7%) or II (14%) (Fig. 2).

Fig. 2 Postoperative improvement in functional status on the basis of New York Heart Association (NYHA) functional class at 3 months.

Chronic nonspecific inflammatory changes were the most common histopathologic findings (33/45; 7.3%). Eight patients (17.8%) had the characteristic histopathologic features of TB, 1 had pericardial involvement with non-Hodgkin's lymphoma, and 1 had pericardial metastatic adenocarcinoma. Of note, 2 patients had normal histologic findings. Microbiologic examinations, including PCR testing in 2 other patients, indicated the presence of TB; the overall TB pericarditis rate, therefore, reached 22.2% in this series (10/45). Table V lists the final diagnoses in our patients, and it shows that TB was the most common known cause of constrictive pericarditis in this patient population.

TABLE V. Final Diagnosis in Patients with Constrictive Pericarditis

Discussion

Constrictive pericarditis is an infrequent cause of a very common condition, heart failure. It is defined as the chronic fibrous thickening of the wall of the pericardial sac, which leads to abnormal diastolic filling.^1,2 The course of the disease is usually slow and the symptoms are nonspecific; consequently, in many cases the symptoms may be present for 12 months or longer before a diagnosis is made.²

The diagnosis of constrictive pericarditis can be a challenging process, but it seems advisable that patients with heart failure but preserved left ventricular function be considered for this diagnosis. Although the right and left ventricular diastolic pressures are equalized in this disease, symptoms of right-sided heart failure are likely to dominate. While it is not definitive, the conventional hallmark of constrictive pericarditis has been the presence of pericardial thickening of more than 3 mm.^2,3

Pericardiectomy is the accepted treatment for constrictive pericarditis; be that as it may, hemodynamic results after surgery range from complete recovery to no benefit.^3–8 These variable results might arise from different approaches to the surgical procedure or from incomplete pericardial resection. The study by Senni and colleagues,³ which had a long follow-up period, reported the persistence of some degree of left ventricular diastolic dysfunction in approximately 40% of the patients, even after total pericardiectomy. The authors posited that this abnormal ventricular compliance might be due to myocardial changes in tandem with incomplete pericardiectomy. DeValeria and colleagues⁶ reported that the myocardial atrophy and fibrosis found at autopsy in constrictive pericarditis patients might indicate some degree of restrictive abnormality of the ventricle after pericardiectomy.

Low-output syndrome was the most common problem during the early postoperative phase in the present study. This low-output state after pericardiectomy might have been due to incomplete pericardiectomy. Nonetheless, the above-mentioned changes, together with postoperative transient interstitial edema, could have had an effect. Entrapment of the heart chambers by insufficiently resected fibrotic epicardium might have caused persistent diastolic dysfunction and a postoperative low-output state. Figure 1 shows that 10 patients needed moderate doses of inotropic agents postoperatively and 3 required high doses; whereas, in the preoperative period, only 7 patients had needed inotropic support. Furthermore, none of the survivors required inotropic agents during the late phase of hospitalization. These changes may appear to be important, but they were not statistically significant (P=0.068). No correlation was found between the use of inotropic agents and the use of CPB (P=0.08). Bozbuga and associates⁹ suggested that low-output syndrome could also be caused by changes in cardiac architecture. In other words, long periods of myocardial compression are likely to contribute to the remodeling of the ventricles and to the weakening of the myocardium in patients who have undergone long periods of symptomatic pericardial constriction. Omoto and colleagues¹⁰ reported that the postoperative low-output state gradually improved in most of their patients.

Most of our patients were operated on through a median sternotomy, which allows excellent access and a better possibility of complete resection. Cardiopulmonary bypass was used in 9 patients; the primary reasons for the use of CPB were severe hemodynamic instability during heart manipulation in 4 patients, concomitant valve replacement in 3, and severe iatrogenic mediastinal bleeding in 2 other patients. Our preference was pericardiectomy without CPB; however, we believe that total pericardiectomy (if possible) using CPB is more advantageous than partial decortication of the pericardium without CPB. Thoracotomy may be preferable in patients who have purulent pericarditis, in which case sternotomy increases the risk of wound infection and sternal dehiscence.¹¹ In our study, the thoracotomy approach was applied in only 2 patients: sternotomy was avoided in 1 patient due to a history of mitral valve replacement and in the other due to cosmetic considerations.

Our patients' NYHA functional status improved considerably after pericardiectomy. Indeed, most of our patients were in NYHA class I or II postoperatively (P <0.001). Except for 3 patients, 1 of whom died in the early postoperative period due to respiratory failure, all patients had better functional status postoperatively. One patient, who had presented in NYHA class III, remained in the same functional class after partial pericardiectomy. Another patient, who was in NYHA class III shortly after pericardiectomy, had shown no evidence of tricuspid regurgitation (TR) in preoperative echocardiographic studies but exhibited moderate-to-severe TR in postoperative studies. By midterm follow-up, this patient had improved with medication and did not need further intervention. Most likely, the TR had been underestimated in preoperative transthoracic echocardiography. Johnson and colleagues¹² reported that post-pericardiectomy TR can be caused by right ventricular dilation. Gongora and coworkers¹³ stated that tricuspid regurgitation may worsen after pericardiectomy.

Buckingham and co-investigators,¹⁴ using transesophageal echocardiography, reported that mitral insufficiency might be seen after pericardiectomy as a result of papillary muscle elongation. In our series, no significant postoperative mitral regurgitation was observed.

In 2 different studies,^15,16 the overall incidence of pericardial calcification detected on chest radiographs ranged between 5% and 27%. Bozbuga and colleagues⁹ found pericardial calcification in 44% of patients with TB pericarditis. In our series, this radiographic finding occurred in 20% of all patients and 30% of patients with TB pericarditis. Statistically, there was no correlation between pericardial calcification and TB (P=0.07). Surprisingly, in the data available to us for review, only a small proportion of our patients had elevated jugular venous pressure (Table I). We would like to emphasize that we were obliged to use medical history sheets. Also, unfortunately, the importance of physical findings in clinical judgment, especially in cardiology, seems to have declined and physical examination has been replaced by routine echocardiographic studies.¹⁷

The overall and in-hospital mortality rates in the present study were 4.4% and 2.2%, respectively. Previously reported data showed in-hospital mortality rates ranging from 4.9% to 16%.6,12,13,18–20 The mortality rate after pericardiectomy in patients with TB pericarditis ranges from 3% to 16%.²¹ The known predictors for post-pericardiectomy death include advanced age, atrial fibrillation, concomitant severe TR, postoperative inotropic support, high pulmonary artery pressure, radiation history, renal failure, low left ventricular ejection fraction, and incomplete pericardiectomy; in the present study, however, neither univariate nor multivariate analysis found a risk factor for early death. This may be due to the small size of our patient population.9,12,13,18

Constrictive pericarditis exhibits a heterogeneous pattern and arises from different causes, depending on the geographic area from which it is reported. In 1999, reports from developed countries indicated that 50% of cases were due to prior pericarditis, cardiac surgery, and radiation therapy, and nearly 30% of cases were idiopathic.²

Tubercular pericarditis is found in approximately 1% of all autopsied cases of TB and in 1% to 2% of patients with pulmonary TB.^9,21 Tuberculosis has been reported to be the cause of nearly 4% of the cases of acute pericarditis, 7% of cardiac tamponade, and 6% of constrictive pericarditis.²² McCaughan and associates²³ reported that active TB was present in 6.1% of their cases of constrictive pericarditis, and Ling and colleagues² reported even a smaller percentage (0.7%). In stark contrast, 38% to 83% of the cases of constrictive pericarditis are still caused by TB in developing countries.^9,12,24

Pericardial involvement in TB patients usually occurs as a result of the lymphatic extension of mycobacterium tuberculosis or of the hematogenous spreading of primary TB infection.²¹ Tubercular pericarditis has 3 clinical presentations: pericardial effusion, constrictive pericarditis, and a combination of the two. The clinical features of TB pericarditis are highly variable, ranging from an absence of symptoms to severe constriction, and the diagnosis is frequently missed on cursory clinical examination.²¹ The diagnosis of TB was made in 22.2% of our patients, which means that despite all the vaccination programs and anti-TB medications in Iran, this disease remains an important cause of constrictive pericarditis. Of note, one third of our TB patients were Afghan immigrants; nonetheless, the incidence of 17% (7 patients) for TB pericarditis in the native Iranian portion of our study group (41 patients) indicates that TB continues to be a frequent source of constrictive pericarditis in our country. We believe that the prevalence of TB pericarditis might be higher than the percentage presented here because of the underdiagnosis of TB infection and the overlap of symptoms between constrictive pericarditis and pulmonary TB. Even in developed countries, it seems that TB pericarditis will not remain a rare disease during the next decade, due to the association between TB and human immunodeficiency viral infection.^21,22

The prevalence of constrictive pericarditis with idiopathic cause was 60% in our series; there is no significant difference between this rate and those reported from the Cleveland Clinic¹⁸ (46%) or from a center in Spain¹ (46.6%). It is of interest, however, that the prevalence of constrictive pericarditis after chest radiation therapy in our patients was considerably lower than those reported by the other 2 centers (Table VI ^1,18). This difference might indicate that our patients with malignant tumors who needed curative or palliative radiation therapy did not achieve similar outcomes, or they may have died before symptoms of constrictive pericarditis became apparent. In addition, the prevalence of post-cardiotomy constrictive pericarditis was 4.4% in our study, compared with 37% reported by the Cleveland Clinic in 2004.¹⁸ However, it is likely that some patients who might otherwise have been in our series did not undergo surgery due to the underdiagnosis of post-cardiotomy constrictive pericarditis or the perceived high surgical risk of reoperation.

TABLE VI. Prevalence of Causal Factors for Constrictive Pericarditis in 3 Representative Treatment Centers in the United States, Spain, and Iran

One of our patients had been exposed to chemical weapons during the Iran–Iraq war. His histopathologic study revealed only nonspecific chronic inflammatory findings. Moreover, our review of the medical literature did not show any distinct relationship between chemical bomb exposure and constrictive pericarditis.

Limitations. Limitations of the present study include its retrospective design, its small number of patients, and its relatively short follow-up period. Therefore, our findings may have shortcomings and may not be conclusive.

Conclusion. Our results show that pericardiectomy remains an effective procedure in the treatment of constrictive pericarditis in that it yields low mortality rates and excellent functional outcomes. Another finding of significance is that, despite intensive vaccination and use of anti-TB drugs in our country, TB is still an important cause of constrictive pericarditis.