Recurrence of pericardial tamponade is relatively common after pericardiocentesis. We evaluated the clinical and procedural predictors of recurrent pericardial tamponade after pericardiocentesis. We included 157 consecutive patients with pericardial tamponade (age 62 ± 18 years, 54% men) who had undergone pericardiocentesis from 2000 to 2007. An intrapericardial catheter was used for prolonged drainage of the pericardial effusion (78% of cases) at the discretion of the operator. The overall recurrence rate 11.8 ± 0.6 months after pericardiocentesis was 20% and the mean interval to recurrence was 1.2 ± 2.1 months. However, patients with extended catheter drainage had a reduced recurrence rate of 12% compared to 52% in patients without extended drainage (p <0.001). In the Cox regression modeling, absence of extended drainage (hazard ratio [HR] 4.1, 95% confidence interval [CI] 1.7 to 10, p = 0.002), incomplete drainage of pericardial effusion (HR 9.7, 95% CI 3.6 to 22.7, p <0.001), loculated effusion (HR 11.1, 95% CI 2.9 to 43, p = 0.001), and malignancy (HR 3.3, 95% CI 1.8 to 10.3, p = 0.037) independently correlated with recurrence at 1 year. In conclusion, extended pericardial drainage after catheter placement is associated with a reduced recurrence of pericardial tamponade after pericardiocentesis.
Percutaneous pericardiocentesis is an established treatment method for the immediate relief of cardiac tamponade. Pericardiocentesis alone is associated with a high risk of recurrent effusion. The extended drainage of a pericardial effusion using a pericardial catheter is safe and reliable. The aim of our study was to evaluate the role of extended drainage on the recurrence of the pericardial tamponade.
Methods
A total of 246 consecutive patients diagnosed with cardiac tamponade by clinical or echocardiographic criteria (2000 to 2007) were reviewed, and 157 were included in our present study. We excluded 65 patients who underwent surgical pericardiotomy because of referral by their physician for surgical drainage. Of these 65 patients, 7 were excluded because of an inability to access the pericardial effusion with the pericardiocentesis needle; 11 underwent surgical exploration for reasons other than effusion recurrence; 5 were excluded because of complications related directly to pericardiocentesis, including right atrial (n = 1) or right ventricular (n = 1) puncture and a blocked or entrapped catheter (n = 3); 1 was excluded because he underwent subsequent pericardectomy for constrictive pericarditis. All patients had 1 year of follow-up.
The biochemical and cellular composition of the pericardial fluid was described without classifying it as exudate or transudate. The institutional review board of Cedars-Sinai Medical Center approved the study (institutional review board approval number Pro00010523).
Standard 2-dimensional echocardiography was used to evaluate the chamber compression or collapse, inferior vena cava size and lack of inspiratory collapse, and the inspiratory variation of the Doppler mitral and tricuspid inflow velocities. A large effusion was defined as a circumferential echo-free space of >1 cm.
Pericardiocentesis was performed as previously described. In those patients receiving extended pericardial drainage, the catheter was left in place until the drainage was ≤100 ml for a 24-hour period, and a follow-up echocardiogram demonstrated the residual effusion to be noncircumferential and <1 cm in size.
Statistical analysis was performed using the Statistical Package for Social Sciences, version 18.0 (SPSS, Chicago, Illinois). Between-group comparisons were performed using the t test or chi-square test. Multivariate Cox regression analysis was performed. Kaplan-Meier analysis was used to estimate survival.
Results
Of the 157 patients, 123 (78%) were treated using pericardiocentesis followed by extended catheter drainage and 34 (22%) underwent pericardiocentesis without extended catheter drainage. The mean duration of catheter drainage was 38 ± 19 hours (range 10 to 96). No significant differences were present in the baseline characteristics between the 2 groups ( Table 1 ).
| Variable | All (n = 157) | Catheter Drainage | p Value | |
|---|---|---|---|---|
| Yes (n = 123) | No (n = 34) | |||
| Age (years) | 62 ± 18 | 61 ± 18 | 63 ± 15 | 0.623 |
| Men | 85 (54%) | 62 (50%) | 23 (68%) | 0.074 |
| Hemodialysis | 16 (10%) | 12 (10%) | 4 (12%) | 0.732 |
| Serum creatinine (mg/dl) | 1.8 ± 2.0 | 1.7 ± 1.8 | 2.1 ± 2.4 | 0.335 |
| Erythrocyte sedimentation rate (mm) | 62 ± 35 | 64 ± 35 | 56 ± 35 | 0.481 |
| Serum lactate dehydrogenase | 337 ± 225 | 345 ± 239 | 286 ± 101 | 0.540 |
| Fluid lactate dehydrogenase | 1,247 ± 1,634 | 1,397 ± 1,753 | 736 ± 1,016 | 0.112 |
| Fluid protein | 5.0 ± 1.2 | 5.0 ± 1.2 | 4.7 ± 0.8 | 0.254 |
| Fluid glucose | 100 ± 70 | 95 ± 67 | 122 ± 84 | 0.170 |
| White blood cell count | 4,570 ± 6,975 | 4,431 ± 5,983 | 5,118 ± 10,188 | 0.784 |
| Hemorrhagic | 35 (22%) | 29 (24%) | 6 (18%) | 0.462 |
The fluid lactate dehydrogenase, protein, glucose, and cell counts were not different between the 2 groups ( Table 1 ). Details on the etiology of the pericardial effusion are listed in Table 2 . The prevalence of malignancy-associated effusion was comparable in the patients with extended catheter drainage versus those without (29% vs 21%, p = 0.359). Cytology of the effusion was positive for malignant cells in 11% of the patients undergoing catheter drainage.
| Variable | All (n = 157) ⁎ | Catheter Drainage | p Value | |
|---|---|---|---|---|
| Yes (n = 123) | No (n = 34) | |||
| After cardiac surgery | 45 (29%) | 37 (30%) | 8 (24%) | 0.439 |
| Associated with malignancy | 42 (29%) | 35 (29%) | 7 (21%) | 0.359 |
| Pericardial fluid cytology positive for malignant cells | 14 (9%) | 14 (11%) | 0 | 0.039 |
| Uremic | 18 (12%) | 12 (10%) | 6 (18%) | 0.201 |
| After viral infection | 12 (8%) | 9 (7%) | 3 (9%) | 0.780 |
| Collagen vascular disease | 3 (2%) | 3 (3%) | 0 | 0.356 |
| After myocardial infarction | 4 (3%) | 3 (2%) | 1 (3%) | 0.869 |
| Traumatic effusion | 4 (3%) | 3 (3%) | 1 (3%) | 0.875 |
| Unknown etiology | 35 (22%) | 25 (20%) | 10 (29%) | 0.261 |
Echocardiographic imaging guidance for the pericardiocentesis procedure was more likely to be used in patients with the extended catheter drainage than in those without extended catheter drainage (98% vs 91%, p = 0.034; Table 3 ).
| Variable | All (n = 157) | Catheter Drainage | p Value | |
|---|---|---|---|---|
| Yes (n = 123) | No (n = 34) | |||
| Echocardiographically guided | 124 (79%) | 102 (83%) | 22 (65%) | 0.021 |
| Fluoroscopy guided | 28 (18%) | 19 (15%) | 9 (27%) | 0.137 |
| Loculated | 3 (2%) | 1 (1%) | 2 (6%) | 0.056 |
| Total volume (ml) | 790 ± 447 | 848 ± 444 | 555 ± 387 | 0.004 |
| Initial volume (ml) | 595 ± 332 | 607 ± 327 | 553 ± 353 | 0.423 |
| Complete drainage | 123 (85%) | 106 (90%) | 17 (65%) | 0.001 |
Comparing the patients with extended pericardial catheter drainage and those without, 29% versus 18% were treated with colchicine (p = 0.175), 24% versus 18% with corticosteroids (p = 0.408), 42% versus 29% with nonsteroidal anti-inflammatory drugs, and 28% versus 21% with aspirin (p = 0.9), respectively. Patients with a malignant effusion were less likely to receive aspirin (7% vs 33%, p = 0.001) than those with a nonmalignant effusion. There was a nonsignificant trend toward less use of nonsteroidal anti-inflammatory drugs (31% vs 42%, p = 0.221) and steroids (17% vs 25%, p = 0.259) in patients with versus without malignancy.
The overall rate of complications related to pericardiocentesis, excluding those requiring immediate surgical intervention, was 5% in the 157 patients. The rate of complications was comparable between those with and without catheter drainage (5% vs 6%, p = 0.778). The complications included supraventricular tachyarrhythmia (n = 2), vasovagal episodes requiring medical attention (n = 3), right ventricular puncture not requiring surgical intervention (n = 1), paroxysmal atrial fibrillation (n = 1), and hypotension with no associated arrhythmia (n = 1). Five patients were excluded after the initial pericardiocentesis because of complications requiring immediate surgical intervention; however, the rate of complications did not differ significantly between the patients with versus those without extended catheter drainage (9% vs 6%, p = 0.605).
At 1 year of follow-up, the overall recurrence rate of pericardial tamponade was 20%, with a mean time to recurrence of 1.2 ± 2.1 months; the mortality rate was 27%. The recurrence rate was significantly lower in patients with than in those without extended catheter drainage (12% vs 52%, p <0.001). However, 33 patients died during follow-up without clinical evidence of pericardial tamponade and were excluded from the analysis evaluating the recurrence of pericardial tamponade. However, these 33 patients were included in the analysis evaluating all-cause mortality and the composite end point of recurrence or death. Mortality was significantly lower in the patients who underwent extended catheter drainage than in those without extended catheter drainage (24% vs 41%, p <0.001). The composite end point of recurrence or death at 1 year of follow up was significantly lower in patients with versus without extended catheter drainage (29% vs 65%, p <0.001). Recurrence-free survival was significantly better in patients with extended catheter drainage than in patients without extended catheter drainage (9.1 ± 0.4 vs 5.2 ± 0.9 months, p <0.001; Figure 1 ) .
The significant predictors of recurrence included a lack of extended pericardial catheter drainage, the initial volume, incomplete drainage, and loculated effusion ( Table 4 ). In the Cox regression analysis, the significant predictors of recurrence included the absence of extended drainage (hazard ratio [HR] 4.1, 95% confidence interval [CI] 1.7 to 10, p = 0.002), incomplete drainage (HR 9.7, 95% CI 3.6 to 22.7, p <0.001), loculated effusion (HR 11.1, 95% CI 2.9 to 43, p = 0.001), and malignancy (HR 3.3, 95% CI 1.8 to 10.3, p = 0.037).
| Variable | Recurrent Effusion ⁎ | p Value | ||
|---|---|---|---|---|
| Yes (n = 25) | No (n = 99) | Univariate | Multivariate | |
| Imaging guidance | 24 (96%) | 96 (97%) | 0.806 | NS |
| Loculated effusion | 3 (12%) | 0 | <0.001 | 0.001 |
| Associated with malignancy | 7 (28%) | 16 (16%) | 0.174 | 0.037 |
| Uremia | 3 (12%) | 9 (9%) | 0.661 | NS |
| High initial volume | 22 (88%) | 55 (56%) | 0.003 | 0.058 |
| Complete drainage | 9 (43%) | 89 (93%) | <0.001 | <0.001 |
| Pericardial catheter | 12 (48%) | 87 (88%) | <0.001 | 0.002 |
| Nonsteroidal anti-inflammatory drugs | 7 (28%) | 48 (49%) | 0.065 | NS |
| Colchicine | 7 (28%) | 30 (30%) | 0.822 | NS |
⁎ Excluding 33 patients who died during follow-up period before having any recurrence.
The 42 patients with malignancy versus those without malignancy were more likely to experience recurrence of the pericardial tamponade or death (62% vs 28%, p <0.001). Of the 42 patients with malignancy, 35 (83%) were treated with extended drainage compared to 88 (77%) of the 115 patients without malignancy (p = 0.359). In patients with malignancy, recurrence of the pericardial tamponade or death occurred in 19 patients (55%) treated using extended pericardial drainage compared to all 7 patients with malignancy treated without extended pericardial drainage (p = 0.023; Figure 2 ). In the subgroup of patients without malignancy, 17 patients (19%) with extended catheter drainage and 15 (56%) without extended catheter drainage experienced effusion recurrence or death (p <0.001; Figure 2 ). Forty-two patients had malignancy associated effusion, including 3 patients with >1 malignancy. Patients with breast cancer and lymphoma had the greatest (33% to 40%) recurrence of pericardial tamponade, and the greatest mortality was observed in patients with lung cancer (92%; Table 5 ).
