Clinical review: Intrapericardial fibrinolysis in management of purulent pericarditis

Purulent pericarditis (PP) is a potentially life-threatening disease. Reported mortality rates are between 20 and 30%. Constrictive pericarditis occurs over the course of PP in at least 3.5% of cases. The frequency of persistent PP (chronic or recurrent purulent pericardial effusion occurring despite drainage and adequate antibiotherapy) is unknown because this entity was not previously classified as a complication of PP. No consensus exists on the optimal management of PP. Nevertheless, the cornerstone of PP management is complete eradication of the focus of infection. In retrospective studies, compared to simple drainage, systematic pericardiectomy provided a prevention of constrictive pericarditis with better clinical outcome. Because of potential morbidity associated with pericardiectomy, intrapericardial fibrinolysis has been proposed as a less invasive method for prevention of persistent PP and constrictive pericarditis. Experimental data demonstrate that fibrin formation, which occurs during the first week of the disease, is an essential step in the evolution to constrictive pericarditis and persistent PP. We reviewed the literature using the MEDLINE database. We evaluated the clinical efficacy, outcome, and complications of pericardial fibrinolysis. Seventy-four cases of fibrinolysis in PP were analysed. Pericarditis of tuberculous origin were excluded. Among the 40 included cases, only two treated by late fibrinolysis encountered failure requiring pericardiectomy. No patient encountered clinical or echocardiographic features of constriction during follow-up. Only one serious complication was described. Despite the lack of definitive evidence, potential benefits of fibrinolysis as a less invasive alternative to surgery in the management of PP seem promising. Early consideration should be given to fibrinolysis in order to prevent both constrictive and persistent PP. Nevertheless, in case of failure of fibrinolysis, pericardiectomy remains the primary option for complete eradication of infection.

plasminogen activator' (tPA), and 'peri carditis' . We reviewed only English-language literature. We found 134 cases includ ing 94 cases reported in one single study, providing pooled results of 60 tuberculous pericarditis patients and 34 PP patients. As patients with tuberculous pericarditis were excluded, only 74 cases were fi nally considered. As the specifi c out come of the 34 PP cases was not available, our outcome analysis only focused on the 40 remaining cases [3].
Information related to modalities of administration, complications, and outcome of fi brinolysis was collected.

Defi nitions and aetiologic classifi cation
PP is defi ned as a neutrophilic pericardial eff usion infected by a bacterial, fungal, or parasitic agent. Tuberculous pericarditis is a lymphocytic eff usion. Classification of PP comprises fi ve aetiologic entities (Table 1) [1,4]. Th e advent of antibiotics has changed epidemiological and aetiopathogenic features. Th e median age of patients has increased from 21 to 49 years [2], whereas the overall frequency of PP has decreased. Before the antibiotic era, PP accounted for 40% of all cases of acute pericarditis and was related to the contiguous spread of lung infections in 72% of cases. Nowadays, pneumonia remains the primary source of infection in only 22% of cases while PP is mainly associated with healthcare and bloodstream infections [1,2,5]. For instance, oesophageal and cardiothoracic surgeries can be responsible for the direct inoculation of the pericardial space whereas immunosuppressive therapy, dialysis, and chemotherapy are relatively recent predisposing conditions [2,4].

Diagnosis
Th e only way to diagnose PP is to obtain pericardial fl uid for microbiological culture and direct microscopic exami nation. In a context of pneumonia, transthoracic echocardiography or computed tomography scan of the chest may disclose complications such as pleural empyema or pericardial eff usion [6,7] (Figure 1). Interestingly, PP may have an insidious presentation without pericardial signs until occurrence of cardiac tamponade [2,4,8,9] (Table 2). Indeed, in a postmortem series by Klacs mann and colleagues, among 55 cases of PP only 10 cases had been suspected and diagnosed ante mortem [1]. Poor diagnostic performance of the condition results in late diagnosis, at an advanced stage, after onset of pericardial adhesions [10].

Complications
PP may lead to cardiac tamponade and septic shock. Th e treatment of these conditions has no specifi city. On the contrary, constrictive pericarditis and persistent PPtwo late-stage complications -imply specifi c considerations and management [1,4].
Constrictive pericarditis is defi ned by the thickening and fusion of the pericardium causing low pericardial com pliance and right heart failure. Haemodynamic compro mise and low cardiac output may ensue because of adiastoly or impaired cardiac fi lling. Pericardial fi brosis is caused by chronic/subacute infl ammation inducing fi bro blast proliferation and collagen deposition. Two large reviews found the same frequency of 3.5% for  constrictive pericarditis during the course of PP [2,11]. In one smaller study with long-term follow-up, however, constrictive pericarditis occurred in 57% of cases, with few cases occurring during the fi rst year. We propose an original defi nition for persistent PP. Th is entity is defi ned by a chronic or recurrent purulent pericardial eff usion occurring despite drainage and adequate antibiotherapy. Th is feature is well described in case reports. Persistence of PP is associated with occurrence of pericardial adhesions and loculation of pericardial eff usion due to the presence of thick fi brin clots preventing complete evacuation of pus through pericardial drains or pericardiostomy [12][13][14][15][16][17]. Purulent discharge persists during several weeks and may lead to death due to septic shock or tamponade. At the stage of persistent PP, pericardiectomy is challenging and associated with high morbidity because of pericardial adherences and areas of fusion. Moreover, patients present with impaired general status [12][13][14][15][16][17]. Interestingly, this entity has not been previously defi ned or classifi ed as a late complication of PP, resulting in a lack of epidemiological data. Finally, because of chronic infl ammation, persistent PP may promote constrictive pericarditis. Nevertheless, these two entities have distinct clinical and pathophysiological features.

Pathophysiology of purulent pericarditis and of its complications
PP shares pathophysiological similarities with empyema. In the case of empyema, the contiguous spread of pneumonia causes pleural mesothelial barrier dysfunction leading to exudative pleural eff usion [18]. Th e second step is bacterial invasion, which attracts leucocytes. Activated leucocytes promote coagulation, resulting in fi brin formation [19]. Fibrin deposits are responsible for pleural adhesion, thickening, and loculation of eff usion, therefore preventing easy drainage and antibiotic diff usion. When the process is prolonged, infl ammation leads to fi brosis due to collagen secretion.
Experimental models argue for a similar pathophysiological process in PP, leading from pericardial infl am mation to pericardial adhesion and fi brosis [20,21]. Th ese studies indicate that pericardial loculation and intra pericardial fi brosis may occur respectively 1 week and 2 weeks after onset of the disease [20,21].
Furthermore, these studies show that fi brin formation is the cornerstone in the pathogenesis of both persistent PP and constrictive pericarditis. Fibrin may thus be a target in the management of PP.

General medico-surgical management
Management of PP requires a combined surgical and medical approach with complete drainage of the eff usion and complementary antibiotherapy. Antibiotic regi men is initially empirical and should target Staphylo coccus aureus. Anaerobes and Gram-negative bacteria should be targeted in case of a digestive origin [22,23]. In recently hospitalised patients, methicillin-resistant Staphylo coccus should be considered [23]. Considering the lack of guidelines on the duration of antibiotherapy in PP and by analogy with empyema, a treatment of at least 3 weeks seems reasonable [24]. Moreover, antibiotherapy should be maintained until removal of the focus of infection, and until improvement of clinical and biological signs.

Surgery
Until 1941 the mortality of PP was estimated to be 100% in untreated cases. Pericardiostomy decreased mortality to about 50%, and to 30% when associated with antibiotics [17]. Eradication of infected pericardial fl uid is the cornerstone of PP therapy, although the way to obtain eradication remains controversial (Table 3) [25]. Guidelines from the European Society of Cardiology make the following statement: 'Frequent irrigation of the pericardial cavity with UK or SK, using large catheters, may liquefy the purulent exudate, but open surgical drainage through subxiphoid pericardiotomy is preferable' [26]. Because of loculations and adhesions, simple evacuation by surgical or percutaneous drainage may not prevent constrictive pericarditis or persistent PP [25,[27][28][29]. To avoid these complications, invasive surgical treat ment has been advocated. In a retrospective series, partial pericardial resection or total pericardiectomy seem to be associated with a better outcome than simple pericardial drainage [4,10,25,27,29] (Table 4). Taken together these studies suggest a curative role of pericardiectomy by achieving complete eradication of the infection and a role in preventing persistent PP and constrictive peri carditis.
Despite these convincing data, physicians are reluctant to perform preventive surgery associated with potential serious complications. Nevertheless, this procedure could be easier early in the disease, before apparition of adhesions. Furthermore, operative mortality is related to the preoperative disability of patients [30,31] and haemodynamic compromise is rarely observed at an early stage. Some authors therefore advocate a systematic early preventive pericardiectomy [32].

Fibrinolysis
Intrapericardial fi brinolysis, a less invasive procedure, has been proposed as an alternative to surgery for PP management. Th e objective of fi brinolysis is to target fi brin formation, to optimise evacuation of a thick fl uid, and therefore to prevent both persistent PP and constrictive pericarditis.

Rationale of intrapericardial fi brinolysis
Fibrinolysis in PP follows the same rationale as that in empyema. In the latter, surgery also has better results than simple drainage. Many authors have attempted to demonstrate the effi cacy of intrapleural fi brinolysis compared with simple drainage for prevention of persistent empyema requiring surgical debridement. Effi cacy of fi brinolysis was demonstrated in a randomised controlled trial (RCT) and a meta-analysis [33,34]. In a double-blind RCT including 44 patients, Diacon and colleagues found a better clinical outcome and fewer referrals to surgery in the SK group [33]. Unexpectedly, the largest RCT testing the effi cacy of SK in empyema did not fi nd any signifi cant diff erence between both groups [35]. Th e group of 208 patients assigned to SK treatment had the same rate of referrals to surgery as the group of 222 patients who received normal saline. Th e population of this study was heterogeneous and the timing of fi brino lysis was late, however, suggesting that fi brinolysis may still be useful in subgroups of younger patients treated earlier.

Evidence of pericardial fi brinolysis effi cacy
Given the similarities between empyema and PP, fi brinolysis has been considered in PP. Th e fi rst cases of successful management of PP with fi brinolysis were reported in 1951 [12,14]. Th ese patients had chronic PP persisting and relapsing despite pericardiostomy and antibiotics. Fibrino lysis was successfully administered at a late stage as a rescue treatment, allowing complete evacuation of purulent material. Forgotten for 30 years, the treatment was rediscovered in 1984 by Bennett [13]. Effi cacy of fi brinolysis was also demonstrated in an experimental model of acute pericarditis in dogs [21]. Fibrinolysis or normal saline was applied through a pericardial drain daily from the third through sixth day after instillation of the irritant mixture. Clinical and histological patterns of constrictive peri carditis were evaluated over a 60-day period. All 11 animals assigned to the normal saline group developed pericardial thickening and adhesions. Among these, nine animals demonstrated clinical features of constrictive pericarditis. One-half of the 10 animals in the UK group did not demonstrate any sign of constriction or any histological changes of the pericardium. Clinical data on fi brinolysis in PP exist. All cases of fi brinolysis in PP available in the English-language literature are summarised in Table 5. Th ere is great heterogeneity in the diff erent protocols used. Among 40 analysable cases, we found two cases of failure requiring pericardiectomy with a favourable outcome. Out of two described deaths, only one was due to PP. No sign of clinical constriction at hospital discharge was observed for the 35 remaining patients (data not provided for three patients). During follow-up, no constriction was found in patients for whom data were available (20 patients on clinical examination and 12 patients on transthoracic echocardiography).
One RCT with long-term follow-up (mean 57 ± 29 months) assessed the effi cacy of fi brinolysis in PP [3]. Unfortunately, the study population included both tuber culous pericarditis (n = 60) and PP (n = 34). Compared with tuberculous pericarditis, PP was 10 times (P <0.0001) more likely to be associated with constrictive pericarditis on transthoracic echocardiography. Fibrinolysis lowered pericardial thickening, and decreased the rate of constrictive pericarditis by 38.3% (hazard ratio = 0.185, P <0.0001). Th e length of evolution of the disease before treatment was also predictive of constrictive pericarditis. Interestingly, the rate of constrictive pericarditis was much higher than in previous reports (3.5% vs. 57% in placebo group), but follow-up was much longer. Constric tion occurred with a mean delay of 11 months. We cannot draw any defi nitive conclusion from this study since no subgroup analysis was performed in the PP population.

Timing of fi brinolysis in PP
Th ere is no consensus on the appropriate timing of fi brino lysis (Table 5). Two main options can be distinguished: primary fi brinolysis performed immediately after drain insertion, and rescue fi brinolysis applied in cases of recurrence or incomplete evacuation of pus. Among published articles, we only found cases of failure when fi brinolysis had been applied late. At this stage,  thick loculations cannot be liquefi ed by fi brinolytic agent [40,42] (Table 5). Furthermore, fi brinolytic agents can only dissolve fi brin, but have no eff ect on fi brosis. Two experimental studies indicate that fi brin infl ux increases during the fi rst week of the disease and that fi brosis may appear after 2 weeks [20,21]. Th ese studies suggest that earlier administration may lead to optimal outcomes. Nevertheless, data regarding optimal timing are solely based on animal models and on a few case reports, limiting the ability to make a defi nitive statement.

Optimal fi brinolytic dose and volume
SK, UK and tPA have all been used in pericardial fi brinolysis. Dose comparison between tPA and UK found an equivalence of 36,000 units UK for 1 mg tPA [49]. Most studies of fi brinolysis in PP used SK. Because of the risk of allergy, however, using UK or tPA has been suggested [50]. Owing to the lack of guidelines, there is large variability in fi brinolytic dosing in the reported cases. UK was used at a dose of 400,000 units in four cases and 200,000 units in another study, diluted in 20 ml normal saline. Th e British Th oracic Society guidelines for the management of pleural infection propose instillation of 250,000 UI SK daily or twice daily or of 100,000 UI UK daily [24]. A dose of 10 mg tPA has been proposed in empyema [51]. Th ere are no defi nitive data on the appropriate dose of these three fi brinolytic agents. Given the diff erent regi mens previously used in PP (Table 5) and the doses proposed by the British Th oracic Society and the Ameri can Th oracic Society for treating empyema,   (Figure 2).

Failure and complications of fi brinolysis
In case of fi brinolysis failure, resulting in incom plete drainage despite three consecutive instillations of fi brinolytic agent associated with adequate antibiotherapy, a surgical approach should be considered. Given the prognosis and the general status of the patient, radical pericardiectomy should only be performed if its benefi ts are believed to outweigh the potential drawbacks. Th ree complications of fi brinolysis can theoretically occur. Even if allergy is a well-known side eff ect of fi brinolysis after repeated instillation of SK, treatment with UK or tPA greatly reduces this risk [50]. Th e second risk is major haemorrhage. Only one case of haemorrhagic tamponade after fi brinolysis has been described [39]. Th is could be a consequence of PP and not directly related to fi brinolysis. Fibrinolysis for empyema is not associated with bleeding events [33,35]. Indeed, fi brinolysis in PP has no systemic eff ect detected on coagulation laboratory tests [28]. Th e third risk is cardiac tamponade, which could occur if the instilled volume is not evacuated by the pericardial drain. Th e pressure-volume curve ( Figure 3) described in the experimental study published by Refsum and colleagues illustrates that the addition of a small volume of liquid may provoke cardiac compression [52]. Th is risk could be increased in the later stage of the disease, characterised by low pericardial compliance secondary to fi brosis. In fact, no tamponade due to intrapericardial administration of fi brinolysis has been described. Th e risk of tamponade could be lowered if the volume of fi brinolytic agent instilled is inferior to the amount of fl uid drained immediately beforehand.
Defi nitive conclusions on the risks and complications of intrapericardial fi brinolysis cannot be made with the evaluation of 74 case reports. Potential benefi ts of fi brinolysis seem to outweigh its apparent low morbidity.

Conclusion
PP managed with simple pericardial drainage may evolve to constrictive and persistent PP. Despite the lack of defi nitive evidence, intrapericardial fi brinolysis may be an alternative to pericardiectomy as a treatment for PP and for prevention of persistent PP and constrictive pericarditis. Clinical and experimental data demonstrate that features of irreversible fi brosis may appear after 2 weeks. Considering that diagnosis of PP is often delayed, early consideration should therefore be given to fi brinolysis to ensure optimal effi cacy. Complications of intrapericardial fi brino lysis have only exceptionally been reported. Physicians should thus not be reluctant to apply this promising technique. Nevertheless, in cases of failure of fi brinolysis, pericardiectomy remains the primary option to consider for complete eradication of infection and prevention of persistent PP and constrictive pericarditis.