Author + information
- Received April 18, 2017
- Revision received June 19, 2017
- Accepted June 22, 2017
- Published online September 13, 2017.
- Kye Hun Kim, MDa,b,
- William R. Miranda, MDa,
- Larry J. Sinak, MDa,
- Faisal F. Syed, MBChBa,
- Rowlens M. Melduni, MDa,
- Raul E. Espinosa, MDa,
- Garvan C. Kane, MDa and
- Jae K. Oh, MDa,∗ ()
- aDepartment of Cardiovascular Diseases, Mayo Clinic, Rochester, Minnesota
- bDepartment of Cardiovascular Medicine, Chonnam National University Hospital, Gwangju, Republic of Korea
- ↵∗Address for correspondence:
Dr. Jae K. Oh, Department of Cardiovascular Diseases, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905.
Objectives This study sought to investigate the incidence, associated findings, and natural history of effusive-constrictive pericarditis (ECP) after pericardiocentesis.
Background ECP is characterized by the coexistence of tense pericardial effusion and constriction of the heart by the visceral pericardium. Echocardiography is currently the main diagnostic tool in the assessment of pericardial disease, but limited data have been published on the incidence and prognosis of ECP diagnosed by echo-Doppler.
Methods A total of 205 consecutive patients undergoing pericardiocentesis at Mayo Clinic, Rochester, Minnesota, were divided into 2 groups (ECP and non-ECP) based on the presence or absence of post-centesis echocardiographic findings of constrictive pericarditis. Clinical, laboratory, and imaging characteristics were compared.
Results ECP was subsequently diagnosed in 33 patients (16%) after pericardiocentesis. Overt clinical cardiac tamponade was present in 52% of ECP patients and 36% of non-ECP patients (p = 0.08). Post-procedure hemopericardium was more frequent in the ECP group (33% vs. 13%; p = 0.003), and a higher percentage of neutrophils and lower percentage of monocytes were noted on pericardial fluid analysis in those patients. Clinical and laboratory findings were otherwise similar. Baseline early diastolic mitral septal annular velocity was significantly higher in the ECP group. Before pericardiocentesis, respiratory variation of mitral inflow velocity, expiratory diastolic flow reversal of hepatic vein, and respirophasic septal shift were significantly more frequent in the ECP group. Fibrinous or loculated effusions were also more frequently observed in the ECP group. Four deaths occurred in the ECP group; all 4 patients had known malignancies. During median follow-up of 3.8 years (interquartile range: 0.5 to 8.3), only 2 patients required pericardiectomy for persistent constrictive features and symptoms.
Conclusions In a large cohort of unselected patients undergoing pericardiocentesis, 16% were found to have ECP. Pre-centesis echocardiographic findings might identify such patients. Long-term prognosis in those patients remains good, and pericardiectomy was rarely required.
Pericardial effusion causes a variety of symptoms depending on its acuity and volume, including dyspnea, chest or abdominal pain, hypotension, and cardiac tamponade, which can be fatal (1,2). Pericardiocentesis is the treatment of choice for patients with symptomatic pericardial effusion. Although symptomatology and hemodynamic abnormalities typically improve dramatically after pericardiocentesis, a subset of patients might fail to show resolution of symptoms or may even worsen after pericardiocentesis. This finding is usually associated with the development of typical features of constrictive pericarditis (CP). This entity has been previously described as effusive-constrictive pericarditis (ECP) (3–5).
ECP is an uncommon clinical syndrome characterized by the coexistence of tense pericardial effusion and constriction of the heart by the visceral pericardium (3–5). The diagnostic hallmark of ECP is the persistence of elevated right atrial pressure measured by invasive hemodynamic assessment after intrapericardial pressure is reduced to a normal level by pericardiocentesis (5). Pericardiectomy was required in more than one-half of the patients with ECP in previous studies. ECP is most likely part of a natural history of pericardial inflammation that occurs with pericardial effusion. Identification of constrictive features in the post-pericardiocentesis setting is important because such patients require closer follow-up.
Although invasive hemodynamic assessment by cardiac catheterization is the gold standard for the diagnosis of ECP, echo-Doppler evaluation is an important diagnostic strategy for various pericardial diseases, including cardiac tamponade and CP (6,7). It is proposed that ECP can be diagnosed post-pericardiocentesis by 2-dimensional and echo-Doppler demonstration of abnormal ventricular septal motion (due to exaggerated ventricular interdependence) and dissociation of intrathoracic and intracardiac pressures, which are the key features of CP (3). However, limited data have been published on the incidence and natural history of ECP diagnosed by echo-Doppler in a large group of patients. Therefore, we sought to investigate the incidence, echo-Doppler findings, and natural history of ECP detected by echocardiography after pericardiocentesis.
From January 2006 to December 2007, pericardiocentesis was performed in a total of 217 patients at Mayo Clinic, Rochester, Minnesota. Among these patients, 12 did not have echocardiographic images available for review. Hence, a total of 205 consecutive patients who underwent echocardiography before and after pericardiocentesis were included in the present study. These patients were divided into 2 groups based on echocardiographic evidence of CP features after pericardiocentesis (ECP and non-ECP groups). The study protocol was approved by the Institutional Review Board of Mayo Clinic.
Baseline and follow-up information was abstracted from clinical notes. Before pericardiocentesis, comprehensive 2-dimensional and echo-Doppler assessments were performed to evaluate the size, location, and hemodynamic effects of the pericardial effusion, if the hemodynamic status of the patient allowed. In hemodynamically unstable patients, echo-Doppler assessments were limited to obtaining essential information, including the location of pericardial effusion and the ideal entry site for pericardiocentesis. Overt clinical cardiac tamponade was defined by a combination of: 1) pulsus paradoxus >10 mm Hg, systemic hypotension (blood pressure <100 mm Hg), or elevated neck veins; 2) presence of hemodynamic instability believed to be secondary to the pericardial effusion; or 3) the need for emergent pericardiocentesis during invasive procedures (8,9).
Echocardiography-guided pericardiocentesis was performed as previously described by our group (10). To ensure complete drainage of the pericardial fluid, a pigtail catheter was introduced into the pericardial space and kept in place until output was <50 cc over a 24-h period. Follow-up comprehensive echo-Doppler studies were performed to assess for the development of ECP within 1 week of pericardiocentesis. The presence of constrictive features was defined by post-pericardiocentesis echo-Doppler findings of inspiratory decrease and expiratory increase of early diastolic mitral inflow velocity (E) >25% accompanied by at least 1 of the following: expiratory diastolic flow reversal of hepatic vein (HV); respirophasic interventricular septal shift; or augmented early diastolic mitral septal annular velocity (e′) and to a level higher than that of the lateral mitral e′.
Fibrinous pericardial effusion was defined as the presence of pericardial effusion with multiple fibrinous strands. Circumferential pericardial effusion was defined as an effusion that encircled the entire heart. Loculated pericardial effusion was defined as an effusion that was located adjacent to 1 or other heart wall or an effusion that was compartmentalized by pericardial adhesion to the heart wall. Pericardial rind was defined as the presence of diffuse pericardial thickening associated with echolucent soft tissues.
The Statistical Package for Social Sciences for Windows, version 13.0 (SPSS, Inc., Chicago, Illinois) was used for statistical analyses. Data are expressed as mean ± SD or median (interquartile range [IQR]; 25th to 75th percentiles) for parametric and nonparametric continuous variables, respectively, and as percentage for categorical data. Chi-square test was used to compare differences in categorical values between the 2 groups. Independent Student t test was used to compare differences in parametric continuous variables, whereas Wilcoxon rank sum test was used for nonparametric continuous variables. Correlations between the variables were established by Pearson correlation. In order to assess the prevalence of ECP in cardiac tamponade related to cardiac surgery/percutaneous interventions versus tamponade unrelated to procedures, we used Poisson regression models with sandwich estimators before and after adjustment for baseline covariates. Values of p < 0.05 was considered statistically significant.
Post-pericardiocentesis echo-Doppler examination was performed after a mean 2.0 ± 1.5 days. Of a total of 205 patients, 33 (16.1%) were found to have ECP (ECP group), whereas 172 (83.9%) did not meet echo-Doppler criteria for constriction (non-ECP group). Overt clinical cardiac tamponade was present in 78 patients (38% of the entire cohort). The presence of clinical cardiac tamponade was similar between groups (p = 0.08).
Baseline clinical characteristics of both groups are summarized in Table 1. Cardiac surgery (28.8%), idiopathic pericarditis (25.4%), procedure-related pericardial effusion (16.1%), and malignancy (11.7%) were the most common causes of pericardial effusion requiring pericardiocentesis. Baseline characteristics were not different between the groups, except for procedure-related pericardial effusion (which is mostly bloody effusion or coagulum tamponade), being significantly more frequent in the ECP group (33.0% vs. 12.8%; p = 0.003). However, using Poisson regression models and adjusting for baseline covariates, cardiac surgery/procedure-related tamponade was not associated with subsequent diagnosis of ECP (risk ratio: 0.90; 95% confidence interval: 0.66 to 1.26).
The results of pericardial fluid analysis are summarized in Table 2. Pericardiocentesis volume was smaller in the ECP group (387.5 ml [IQR: 308.8 to 547.5 ml] vs. 500 ml [IQR: 365 to 700 ml]; p = 0.046). Although there was no significant difference in total leukocyte count between the 2 groups (1,950.0 per mm3 [IQR: 787.5 to 3,662.5 per mm3] vs. 1,175.0 per mm3 [IQR: 200.0 to 3,487.5 per mm3]; p = 0.18), the percentage of neutrophils was significantly higher (50.0% [IQR: 25.0% to 65.5%] vs. 19.5% [IQR: 5.0% to 42.3%]; p = 0.004), and the percentage of monocytes was significantly lower (11.0% [IQR: 4.0% to 19.5%] vs. 25.0% [IQR: 9.5% to 48.0%]; p = 0.007) in ECP compared to non-ECP patients. No differences in other pericardial fluid parameters were identified.
Echocardiographic features pre- and post-pericardiocentesis
Pre-pericardiocentesis echo-Doppler findings are summarized in Table 3. Figure 1 illustrates echo-Doppler findings pre- and post-pericardiocentesis in a patient with ECP. Figure 2 shows post-pericardiocentesis respirophasic shift, mitral inflow, and tissue Doppler findings in a different patient with ECP (Online Videos 1 and 2). Mean medial mitral e′ velocity was higher in the ECP group (8.9 ± 2.5 vs. 6.9 ± 2.4; p < 0.001), and respirophasic interventricular septal shift was more frequently observed in ECP patients (21.2% vs. 1.2%; p < 0.001). Mitral inflow variation was seen in 89.3% of ECP patients and 62.3% of non-ECP patients (p = 0.006). Expiratory HV flow reversals were also commonly encountered in the ECP group (48.0% vs. 22.5%; p = 0.009). There were no differences in mitral inflow deceleration time or in expiratory E or inspiratory E- or A-wave velocities on transmitral Doppler. Fibrinous and loculated pericardial effusions were also more common in the ECP group.
Table 4 lists post-pericardiocentesis echocardiographic findings. By definition, mitral inflow variation was present in all of the ECP patients, as opposed to only 1.2% of non-ECP patients. Respirophasic septal shift was present in 97% of ECP patients, and dilation of the inferior vena cava was present in all; those features were present in 21.5% and 41.3% of non-ECP patients, respectively (p < 0.001 for both). Although no significant differences in E- and A-wave velocities were seen, mitral inflow deceleration time was shorter in the ECP group. Diffuse pericardial thickening was present in 72.7% of ECP patients but in only 19.2% of non-ECP patients (p < 0.001).
Additional imaging and cardiac catheterization data in patients with ECP
Four ECP patients underwent computed tomography (CT) scanning within 7 days after pericardiocentesis. Only 1 study was performed to assess the pericardium, and mild pericardial thickening was present in this patient. Ten patients underwent CT within a week before pericardiocentesis. Four of those were dedicated cardiac CTs (performed before cardiac ablation); all patients had normal pericardial thickness without any evidence of CP. None of the patients underwent cardiac magnetic resonance within 7 days pre- or post-pericardiocentesis.
One ECP patient underwent cardiac catheterization the day after pericardiocentesis. Right atrial pressure was 19 mm Hg, and hemodynamic findings were consistent with constrictive physiology.
Clinical outcomes of post-pericardiocentesis constrictive features
Of the 33 patients with constrictive features, 16 were treated with steroids (n = 3) or nonsteroidal anti-inflammatory therapy (n = 13). Colchicine was prescribed to 6 patients (as monotherapy in 3). During follow-up (median 3.8 years [IQR: 0.5 to 8.3 years]), 4 patients died early in their clinical course; all of them had known malignancies. Six patients (18%) were rehospitalized with progressive shortness of breath or symptoms of heart failure (median 33 days [IQR: 29 to 78 days] after initial pericardiocentesis). Follow-up echocardiography (median 134 days [IQR: 36.5 to 395.5 days]) was available for 26 of 33 patients (79%) and showed resolution of constrictive features in 24 of them. Two patients showed persistent constrictive physiology and symptoms despite anti-inflammatory therapy and underwent successful pericardiectomy. In both cases, pathology revealed pericardial thickening, and microscopic findings were consistent with CP.
We present herein the incidence of ECP after pericardiocentesis and its long-term prognosis in a large group of patients seen at a tertiary care center. Among several important observations, our data suggest that: 1) echocardiographic features of CP are common post-pericardiocentesis; and 2) ECP resolves in most patients but may require pericardiectomy in a small number of cases.
Features of CP after pericardiocentesis were observed in 16% of our cohort. This number is significantly higher than the 1% to 2% reported in unselected series of patients with pericarditis (4,11), and it is still higher when compared to the subset of patients presenting with cardiac tamponade (8%) (4). However, it is lower than the rates of ECP observed in patients requiring pericardiectomy for CP (24% prevalence) (12). According to a recent systematic review including a total of 642 patients with pericarditis/pericardial effusions, the prevalence of ECP varied between 2.4% and 14.8% (3).
There are several different explanations for the wide variation in the incidence of ECP reported in the literature. First, the methodology used to diagnose ECP varied substantially between groups (4,9,11,13,14). Some studies included simultaneous cardiac catheterization in all patients undergoing pericardiocentesis; in other studies, the diagnosis was based on echocardiography or a combination of both modalities. Moreover, the echocardiographic assessment of CP has evolved tremendously over the years. Currently, echo-Doppler allows for the diagnosis of constrictive physiology with greater sensitivity and specificity than ever before (15). Thus, cardiac catheterization is now reserved for patients whose noninvasive evaluation is inconclusive. Defining the “expected” incidence of constrictive features post-pericardiocentesis by echocardiography and its associated findings is of clinical importance because it might guide follow-up and prevent the need for further testing or procedures.
The incidence of ECP has also varied according to the studied populations and associated etiologies. For example, rates of ECP in tuberculous pericarditis have been reported to be as high as 38% (16). The risk of requiring pericardiectomy in ECP also appears to be directly related to the underlying cause, varying from 50% to 73% according to the series (3). In contrast, Sagrista-Sauleda et al. (17) reported resolution of constrictive features in all 16 patients diagnosed with transient constriction; those patients found to have constrictive features in the convalescent phase of acute pericarditis. Because of differences in methodology and underlying etiologies, comparison between studies is difficult, and the prognosis of ECP/transient constriction is still poorly understood. Moreover, studies published more than a decade ago involved imaging during an era when the awareness of “uncommon” patterns of CP was not fully appreciated, and therapeutic options were more limited and less standardized (18). Thus, it is possible pericardiectomy was performed in patients with transient or reversible constriction. It is noteworthy that idiopathic and post-operative cases accounted for >40% of our ECP patients. In addition, one-third of ECP cases were seen in procedure-related effusions (vs. 13% of non-ECP cases). Although cardiac surgery- or procedure-related effusions were not associated with higher risk of ECP, our results should be extrapolated with caution to populations that differ from ours.
Interestingly, overt clinical cardiac tamponade was present is only one-half of our patients with ECP. Although the prevalence of tamponade physiology by cardiac catheterization could have been higher, our results would be in agreement with the observations of Ntsekhe et al. (16). In their study, cardiac tamponade diagnosed by right heart catheterization was present in 53% of ECP cases, a rate similar to that of patients without ECP (56%). Although originally described in the setting of cardiac tamponade, those results support that the distinction between ECP and transient CP is essentially academic. Both entities are part of a clinical spectrum of patients presenting with pericardial inflammation, less compliant pericardium, and pericardial effusion. The hypothesis of inflammation playing a significant role in the pathophysiology of ECP (5) is also supported by our data, as a higher proportion of acute inflammatory cells is seen in the pericardial fluid of ECP patients. Pericardial thickening by 2-dimensional echocardiography was also more frequent in that group, and the reversibility of constriction could be predicted by intense pericardial inflammation noted on cardiac magnetic resonance and by increased inflammatory biomarkers (19). Therefore, our finding suggests that patients who undergo pericardiocentesis are best served by a period of anti-inflammatory treatment to shorten the period of ECP or perhaps even lessen the possibility of chronic CP. It should be noted that one-fifth of ECP patients were rehospitalized because of dyspnea or heart failure within 3 months. Although the long-term prognosis is good, those patients might require closer follow-up shortly after their index event.
Our results also suggest that patients with ECP might have distinct echo-Doppler features even before pericardiocentesis compared to patients with effusive pericarditis. Higher mitral medial e′ velocity and respirophasic septal shift, features typically encountered in patients with CP (15), were more frequent in the ECP group. This is also the most likely explanation for the higher prevalence of mitral inflow variation and HV flow reversal in the ECP group. Although those findings are seen in tamponade physiology (8), the prevalence of clinical cardiac tamponade was similar between groups. To our knowledge, pre-pericardiocentesis echo-Doppler findings in patients with ECP have never been reported. Whether those features can predict post-pericardiocentesis findings and long-term prognosis requires further investigation.
Our study represents the largest to assess the incidence of constrictive features/ECP in patients undergoing pericardiocentesis in the United States. Although multiple studies have described the incidence of ECP in developing countries, the epidemiology of patients presenting with acute pericarditis/pericardial effusions is very different in North America and Europe than in other parts of the world (2). In addition, our study is the first to include a comprehensive echo-Doppler assessment of constrictive features post-pericardiocentesis. This is clinically important because echocardiography currently is the main diagnostic tool for pericardial diseases (1), and the incidence of echo-Doppler features of CP post-centesis in unselected patients has not been described. Our data suggest that the long-term prognosis in such patients is good, and that observation and medical management rather than early pericardiectomy are the preferred therapy. However, our favorable results need to be reproduced in other large-scale studies.
This is a single-center, retrospective study. Although our institution is a tertiary center, we included consecutive patients in order to minimize selection bias. However, referral bias might have contributed to the higher prevalence of ECP observed in our study compared to others. Simultaneous cardiac catheterization during pericardiocentesis is not routinely performed in our practice, and the incidence of ECP based on invasive versus noninvasive assessment might have differed. Further studies comparing invasive hemodynamic criteria to imaging criteria for ECP are required to address this question. In addition, a small proportion of patients in the non-ECP group had some but not all features of constriction, and whether these patients would have been diagnosed with ECP using invasive hemodynamic assessment is unknown.
The results of our study showed that evidence of ECP post-pericardiocentesis is common and that CP features usually resolve either spontaneously or with medical management. These observations suggest a reversible, inflammatory cause for the hemodynamic abnormality and support a conservative approach to management, reserving pericardiectomy for patients refractory to adequate anti-inflammatory therapy.
COMPETENCY IN MEDICAL KNOWLEDGE: Echocardiographic features of CP are common after pericardiocentesis, but the long-term prognosis is good. Pericardiectomy was necessary in only a small number of patients. Pre-pericardiocentesis, high mitral medial e′ velocities and respirophasic septal shift were seen more frequently in the ECP group, suggesting that these patients have distinct echo-Doppler features even before pericardiocentesis is performed.
TRANSLATIONAL OUTLOOK: Data regarding the natural history of ECP are limited, and our results need to be reproduced in other populations. In addition, studies correlating echocardiographic findings and invasive hemodynamics post-pericardiocentesis are needed to allow better understanding of the pathophysiology and prognosis of ECP.
For supplemental videos and their legends, please see the online version of this article.
All authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- constrictive pericarditis
- computed tomography
- mitral inflow early diastolic velocity
- early diastolic mitral septal annular velocity
- effusive-constrictive pericarditis
- hepatic vein
- Received April 18, 2017.
- Revision received June 19, 2017.
- Accepted June 22, 2017.
- 2017 American College of Cardiology Foundation
- Adler Y.,
- Charron P.,
- Imazio M.,
- et al.
- Nugue O.,
- Millaire A.,
- Porte H.,
- et al.
- George S.,
- Salama A.L.,
- Uthaman B.,
- Cherian G.
- Sagrista-Sauleda J.
- Feng D.,
- Glockner J.,
- Kim K.,
- et al.