Applied Radiology

Percutaneous imaging-guided drainage of complicated fluid collections-abscesses, empyemas, hematomas, lymphoceles, and the like-has become a well established practice. Ideally, these collections are drained quickly, safely, and completely, with as few ancillary tube manipulations or additional procedures as possible. Recently, the concept of utilizing intracavitary fibrinolytic agents to hasten and improve drainage of such collections has gained interest and favor. In this article, we will review the use of such therapy in a "question and answer" format. One caveat is necessary before embarking upon this discussion. Although we, and others, are enthusiastic about the role of fibrinolytic agents in complicated fluid drainages, intellectual honesty mandates that we acknowledge that there is a great deal of work to be done in this area. In particular, randomized, prospective studies evaluating the efficacy, risks, and costs of intracavitary fibrinolytic therapy relative to standard drainage protocols are lacking. We write this, in part, in the hope of spurring more rigorous investigation as to the appropriate role and optimal algorithm for the use of this therapy.

What is the rationale for using fibrinolytic therapy in complicated fluid collections?

Let us first consider infected collections. Deposition of fibrin appears to play an major role in the host-defense to infection.1-8 The formation of fibrin is initiated via an early inflammatory response in which leukocytes that have migrated to the site of infection release permeability factors, in turn causing protein-rich exudative fluid to spill from blood vessels into the affected area. Conversion of fibrinogen to fibrin then occurs and is promoted by several mechanisms; in the peritoneal cavity, for example, this has been shown to result from release of tissue thromboplastin by mesothelial cells and stimulated macrophages, as well as from reduced peritoneal fibrinolytic activity.7,9 Fibrin walls off the infection, preventing its spread by causing adherence of adjacent pleural or peritoneal surfaces; the fibrin matrix, to a variable degree, also binds and traps microorganisms. By limiting the spread of infection and delaying systemic sepsis, this host response is beneficial. Unfortunately, from the host standpoint, it also has a downside: While fibrin protects the host against spread of microorganisms, it also protects the microorganisms from host defenses, promoting abscess or empyema formation. Further, and from the standpoint of the interventional radiologist, fibrin increases the viscosity of the collection and may result in multiple fibrinous loculations, making the collection progressively more difficult to drain percutaneously. Series assessing the results of percutaneous abscess and empyema drainages have repeatedly shown diminished success with complex, multilocular collections or collections with thick, tenacious contents.10-13 If fibrinous material could be eliminated in conjunction with drainage, one would anticipate that improvements in efficiency and success of drainage could result. Fibrinolytic agents such as urokinase (UK) and streptokinase (SK) promote conversion of plasminogen to its active proteolytic form, plasmin; plasmin in turn lyses fibrin. Installation of lytic agents into complex collections would thus seem a reasonable adjunct to drainage of such collections.

Are fibrinolytic agents enzymatically active within complicated fluid collections?

Lahorra et al, in their study of intracavitary UK for percutaneous abscess drainage, measured fibrinogen and fibrin split products in drained abscess fluid before and after treatment with intracavitary UK.8 They found that fibrinogen levels decreased, while fibrin degradation products increased sharply then decreased to baseline pre-treatment levels after administration of UK. Similarly, Moulton et al noted that measurable plasminogen in pleural fluid before treatment with UK decreased to undetectable amounts after treatment.14 Such data suggest that the enzymatic activity of UK is maintained in abscesses and complicated pleural collections; other lytic agents have not been tested in this manner.

Do fibrinolytic agents decrease the viscosity of infected fluid collections?

Early studies using thrombolytic agents in empyemas and hemothoraces commented frequently on qualitative "thinning" of the drained fluid; however, there was little in the way of quantitative evaluation. One exception is the study by Tillett and colleagues, which examined (without mention of methodology) the viscosity of fluid from a tuberculous empyema before and after injection of streptokinase/streptodornase.15 They found that the lytic agent decreased the viscosity of the fluid from 3,000 times that of water to 33 times that of water within 4 hours, and to 13 times that of water by the next day.

Park and colleagues tested the hypothesis that a fibrinolytic agent-in their study, UK-could diminish the viscosity of complex collections in an in vitro study.16 By mixing UK with purulent fluid, they found (based on measurement with a viscometer) that the viscosity of the fluid decreased by 23% compared to purulent fluid alone. Further, they noted that the addition of urokinase to purulent fluid significantly shortened drainage time of 10 ml of fluid through seven different drainage catheters ranging in diameter from

6-French through 18-French.

You've given a rationale for use of fibrinolytic agents in infected fluid collections. Is there any rationale for their use in noninfected complex fluid collections?

Certainly there is good reason to believe that noninfected hematomas and hemothoraces should respond to fibrinolytic therapy; clinical data (as detailed below) support this. Whether or not they should be treated is another question, as many such hematomas will resolve without therapy. On the other hand, there may be some justification in using lytic agents fairly early in the setting of noninfected hemothoraces to prevent organization and fibrosis of the collection and subsequent lung entrapment, or to prevent empyema formation.14,17,18 Another argument for drainage of noninfected hematomas or other complex collections is to treat adverse clinical sequelae-for example, hypoxemia and atelectasis-which develop because of mass effect.14

Malignant pleural effusions can be associated with a pleural inflammatory response and formation of fibrin septae.19,20 As a result, one might anticipate that lytic agents could facilitate drainage in this setting as well, potentially enabling or hastening attempts at pleurodesis. There are, however, only limited clinical data in this regard.21,22 More confident conclusions with respect to these or other noninfected complex collections (such as lymphoceles) will require additional studies.

Does clinical data support the use of fibrinolytic therapy in complicated fluid collections?

Interestingly, the first clinical use of lytic agents as adjuncts to complex fluid drainage was described nearly half a century ago, when Tillett, Sherry, and other investigators began using streptokinase (SK) and streptodornase (SD) in the pleural space for treatment of empyema and hemothorax.15,23-27 Evidence for efficacy of the treatment in these studies included response of some infections (including chronic infections) which had previously failed other standard therapies, qualitative thinning of drained material, increases (often dramatic) in the amount of fluid drained, decreases in quantitative bacterial counts from infected collections, and the appearance of fresh granulation tissue in instances in which wounds were visible.

Although the results of these early studies were generally favorable, particularly if the agents were used early, many patients developed fever, malaise, and leukocytosis. Interest in the technique dwindled, and there was a hiatus of 18 years (between 1959 and 1977) during which no reports of this use of fibrinolytic agents appeared in the literature. In 1977, Bergh et al revisited enzymatic débridement of the pleural space and reported lung reexpansion in 30 of 38 cases treated with SK instillation after conventional drainage had proved ineffective.28 They encountered only minor, self-limited increases in temperature in some of their patients, and implied that the decreased incidence of side effects resulted from their use of purified SK rather than the previously used combination of SK-SD.

Several other series and case reports of fibrinolytic débridement of the pleural space utilizing SK, SK-SD, and UK have since appeared.14,18,29-35 Generally, all have been enthusiastic about the use of fibrinolytic agents for this purpose, with success rates in series ranging from 44 to 100%. Lack of success has been related, at least in part, to the age of the pleural collection, with older collections responding less well. This is postulated to result from development of organized fibrous tissue in the pleural space which will not respond to lytic therapy.18 As with earlier studies, claims as to the efficacy of drainage in these reports are based primarily upon improved drainage in previously refractory or poorly draining collections, and not on direct comparison to any control group.

In the largest series to date, that of Moulton and colleagues, adjunctive UK was utilized in 98 of 118 pleural drainages, which included 79 empyemas, 27 sterile loculated parapneumonic effusions, 10 sterile hemothoraces, and 2 sterile postoperative exudative effusions.18 Their criteria for initiating UK was the presence of significant fluid on follow-up imaging, providing that drainage catheter position was satisfactory. Using UK concentrations of 1000 U/ml or 2500 U/ml, administered in volumes of 20 to 240 ml (depending on the size of the collection) and a range of 2 to 17 instillations per case (average 4.9), the authors reported an overall success rate of 94%. The mean dose of UK in this series was 466,000 U (slightly greater for hemothoraces than nonhemorrhagic effusions) and the mean duration of chest tube drainage was approximately 6 days. In arguing for the use of adjunctive UK, the authors point out not only the high success rates of drainage (with avoidance of the need for open drainage or decortication) and the often dramatic increases in fluid output following initiation of therapy, but they also note the shorter mean duration of drainage in comparison to standard or imaging-guided chest tube drainage.

In 1987, Vogelzang and colleagues reported on the first use of lytic therapy in drainage of infected extrathoracic collections.36 They reported improvement in drainage using streptokinase in one instance and urokinase in two instances of infected postsurgical hematomas, with improvement judged by virtue of significant increases in tube output and in resolution of the collections following an initial trial of standard drainage. In another study of abdominal abscesses, designed to assess safety of adjunctive lytic therapy, Lahorra et al reported high success rates particularly for their two higher dosing regimens. They also reported successful drainage in all eight abscesses in which septations were identified by imaging.8 While the lack of a control group precluded confident assessment of efficacy, the authors stated they were "cautiously enthusiastic" regarding the potential value of adjunctive UK in abdominal abscess drainage.

Details of other clinical series are provided in answers to other questions below.

What is the best fibrinolytic agent to use?

While no direct comparison studies have been done, there has been an increasing tendency in recent years to use UK. The advantage of UK over SK-SD or SK alone is that it produces no antibody response. Antibodies to SK, typically the result of prior exposure to streptococcal infections, are quite common and affect a large percentage of most treatment populations. Such antibodies are felt to have accounted for the reactions described previously in response to SK-SD (although, as noted, SK alone seems to have fewer problems). Antibodies also may be associated with deactivation of SK and, thus, diminished fibrinolytic efficacy.18,28 On the other hand, UK, a naturally occurring human enzyme, has not been associated with allergic reactions when used in the pleural space.

Other fibrinolytic agents, such as tissue plasminogen activator, have not been used for enzymatic débridement of complicated fluid collections. As an aside, there has been interest in the use of another proteolytic agent, N-acetylcysteine, in a manner similar to that of the fibrinolytic agents. However, it has been shown that N-acetylcysteine has little activity at the pH values found within abscesses in vitro;37 there has been little recent interest in this agent.

When should fibrinolytic agents be used in the pleural space?

Empyema, defined by the presence of grossly purulent fluid in the pleural space, may arise by several mechanisms.2 Frequently, it develops in conjunction with pneumonia or other infectious processes of the lung, but-as in the case of post-surgical or post-traumatic empyema-it need not always do so.

When pleural effusions are associated with pneumonia, as is often the case, they are termed "parapneumonic effusions."2 When unchecked by host defenses or appropriate therapy, parapneumonic effusions evolve into increasingly complex collections and, in a series of roughly defined stages, become empyemas. In Stage 1 (exudative stage), the parapneumonic effusion consists of thin, sterile fluid with a relatively low white blood cell count. In Stage 2 (fibrinopurulent stage), bacteria and increasing numbers of white blood cells invade the pleural space. Additionally, fibrin begins to form within the effusion, creating septations within the effusion and a layer of fibrin along the pleural surfaces. Stage 3 (organization stage) is the end result of a chronically infected pleural space, in which a thick peel of organized fibrous tissue forms along the pleural surfaces.

Stage 1 parapneumonic effusion will resolve without the need for chest tube drainage, assuming the parenchymal infection responds to host defenses and/or appropriate antibiotic therapy. As the parapneumonic effusion becomes more complicated, chest tube drainage becomes necessary for optimal therapy. The efficacy of chest tube drainage decreases with progressive fibrin deposition within the pleural space. Fibrinolytic agents would logically be most beneficial in Stage 2, after fibrin deposition occurs but before an undrainable and unlysable thick fibrotic peel or rind develops. Once the latter occurs, management turns to more aggressive and/or prolonged therapies, such as open drainage or pleural decortication.

Three types of data can help determine when to initiate fibrinolytic therapy: pleural fluid analysis, imaging studies, and response to chest tube drainage. Light has suggested a treatment algorithm based on a further subclassification of parapneumonic effusions and empyema.2 In his schema, tube thoracostomy is advised once a para-pneumonic effusion has become complicated, defined by pleural fluid analysis which indicates pH less than 7.00, glucose less than 40 mg/dl, or a positive Gram stain or culture. Adjunctive fibrinolytic agents are recommended when there is evidence that such complicated effusions are multiloculated. This can be determined by imaging studies, in particular ultrasonography which readily demonstrates septations within the pleural collection. On the other hand, there is recent evidence that heavily septated collections-the "honeycomb" pattern described by Park and colleagues33-may not respond well to fibrinolysis (figure 1). Computed tomography (CT) does not demonstrate these septations as well as ultrasonography, although pockets of air failing to rise to the least dependent portion of the pleural space, if present on CT, indicate that loculations are present (figure 2). However, CT can give an excellent overall global assessment of the pleural space, and is particularly useful in two settings:

1) prior to drainage to establish an overview of the pleural collection, identify any areas of loculation or the presence and degree of pleural thickening, and help determine locations for optimal tube placement, and 2) to determine reasons for incomplete response to initial tube drainage. In the latter case, if CT shows good tube positioning and no evidence of locules remote from the tube, fibrinolytic agents are recommended (figures 2,3). Assessment of pleural thickening by CT has recently been reported to be helpful in predicting response of loculated pleural collections to fibrinolytic therapy. Park and colleagues found that the effectiveness of UK varied inversely with parietal pleural thickness in 31 treated patients, reporting no effective drainages in 3 patients with a thickness of 5 mm or greater.33

What are the indications for the use of fibrinolytic agents in the abdomen and other areas?

The indications for fibrinolytic therapy in collections outside of the pleural space are less well defined. In our own practice, we do not use such agents frequently, but will consider them in two settings: in cases of infected hematomas-as in the post-surgical collections described earlier in the cases reported by Vogelzang et al36-or in cases in which drainage from an infected collection is poor despite a patent and seemingly well positioned drainage catheter.

A recent abstract by Ryan et al described a similar selective use of UK in abdominal fluid collections.38 Out of 685 patients who underwent percutaneous drainage of abdominal fluid collections over a period of 6 years, they initiated intracavitary fibrinolytic therapy in 15 patients (2.2%); their criteria for adjunctive use of UK included incomplete drainage of the collection on follow-up imaging and/or the presence of a hematoma. They found some response (partial or complete resolution) in 86.7% of the cases and suggested, in particular, that UK is useful in patients with infected post-surgical hematomas.

A potential, but as yet clinically untested application of urokinase in abscess drainage is in the setting of infected prosthetic vascular grafts. Infection of such grafts is an exceedingly serious complication which typically requires graft excision. It is believed that an important etiologic factor in infection of grafts and other prosthetic materials is the development of a "biofilm" which effectively shields the contaminating bacteria from antibiotics and host defenses. By assisting in the breakdown of this biofilm, as well as by degrading fibrin within any associated perigraft abscess, urokinase might permit sterilization and salvage of infected grafts. Nakamoto and colleagues, in a paper which won the SCVIR Young Investigator Award, tested this hypothesis in an animal model.39 They found that antibiotics and intraabscess urokinase were synergistic in their ability to sterilize infected graft implants in hamsters.

One other unusual application of fibrinolytic therapy in drainage was described in a case report by Stempel and Vogelzang.40 They reported the use of UK in conjunction with percutaneous cholecystostomy to treat hemorrhagic cholecystitis; this was associated with hemobilia and resultant obstructive jaundice. Lytic therapy in this case produced rapid clearing of blood clots from the biliary tree, facilitating resumption of normal bile flow.

Are there any risks or contraindications associated with intracavitary administration of fibrinolytic agents?

The most common, as well as the most serious complications associated with the use of fibrinolytic agents in the vascular system are hemorrhagic in nature, either at vascular puncture sites or in remote locations.41 It is not surprising, then, that most commentary and investigation regarding the safety of such agents in complicated extravascular fluid collections have focused on hemorrhagic complications or effects on systemic coagulation parameters. Despite this, however, significant hemorrhagic complications have been exceedingly rare in relation to the use of extravascular fibrinolysis. We are aware of one case report in which a patient, following intrapleural administration of 500,000 units of SK, developed hematuria, epistaxis, and bleeding from chest, nasogastric, and endotracheal tubes; this was associated with increased prothrombin and partial thromboplastin time, as well as a drop in hematocrit from 30.3 to 22.7, which necessitated transfusion.42 The patient had chest tubes placed for a post-decortication pleural collection associated with persistent bronchopleural fistula, and it was speculated that vascular erosion and/or the bronchopleural fistula may have predisposed the patient to systemic absorption of the lytic agent.

Apart from the above report, systemic coagulation parameters have not been adversely affected in those reports in which they were assessed.8,14,29,36 Some authors have found that pleural drainage fluid has become hemorrhagic during the course of lytic therapy, but not to an extent which has been clinically apparent.18 Finally, Bergh et al did report a decrease in hemoglobin in 7 of 38 patients treated with intrapleural streptokinase, but did not describe any drop to result in serious adverse complications, and further acknowledged that it was difficult to assess whether the drop was related to underlying disease or the lytic therapy.28

In 1993, Lahorra and colleagues reported an FDA approved phase I safety study of the use of UK in abscess drainage.8 In this study, 31 abscesses in 26 consecutive patients were treated with UK at a concentration of 5000 IU/ml. Three dosing regimens were used, with each dose consisting of 1000 IU, 2500 IU, or 5000 IU of UK per centimeter of greatest abscess diameter; UK was administered every 8 hours for 3 days. No hemorrhagic complications occurred in this study; indeed, the only adverse effect encountered was discomfort during UK injection in one patient, not severe enough to discontinue treatment. Additionally, there were no significant changes in hematocrit, prothrombin time or partial thromboplastin time, platelet count, or serum fibrinogen levels; while there was an increase in fibrin degradation products at 24 hours after initiation of therapy, this elevation was transient and was not statistically significant.

Despite such data, it may be reasonable to exclude certain patients from intracavitary lytic therapy based on theoretical considerations. For example, Lahorra and colleagues excluded patients with a history of CNS bleeding or with potentially hemorrhagic CNS lesions, known coagulopathy or thrombocytopenia, splenic abscess (because of the highly vascular nature of the spleen), intrapancreatic abscess (because of concern regarding the presence of inflammatory pseudo-aneurysms associated with pancreatitis), interloop abscess (because of uncertainty regarding the effect of UK on closure of enteric fistulae), or pregnancy; Nakamota and Haaga have elaborated very similar criteria, also including known hypersensitivity to UK.1,8 With regard to the latter, no hypersensitivity or idiosyncratic reactions to intracavitary UK have been described: neither the fevers and malaise reported in early experience with streptokinase-streptodornase combinations nor the rigors and other reactions described during intravascular thrombolysis with UK.41

An interesting, but as yet unanswered question is how soon after trauma or surgery can fibrinolytic agents be safely administered. Moulton et al specifically avoided collections within 3 days of a known bleed in one series, but added that one may need to balance the theoretical risk of hemorrhagic complications against the risk that delay in therapy may result in pleural fibrosis.14 In the larger series of thoracic drainages reported by Moulton and colleagues, no hemorrhagic complications occurred in treating 10 hemothoraces between 5 and 165 days of the initial bleed, or in treating two post-surgical collections at 6 and 30 days following operation.18 In commenting on this issue, these investigators noted that other authors have described lytic therapy within 1 to 3 days post trauma or post surgery without encountering problems.23,25

One other setting in which fibrinolytic therapy is felt to be contraindicated is bronchopleural fistula; lytic agents could theoretically prevent healing of a bronchopleural fistula, or cause one to reopen. While the former remains a concern, evidence suggests that the risk of opening a recently closed fistula is small.18,32

How is urokinase administered in complicated fluid collections?

Reported dosing regimens for intracavitary fibrinolytic agents have been variable, without any clear consensus or evidence favoring one protocol over another in the literature on this subject. We will confine ourselves to the the regimens involving UK.

Questions to be answered regarding dosing include the diluent to be used, the concentration of UK, the volume administered, how long the agent is to be kept in contact with the collection, how the fluid is then to be drained, how often the UK is administered, how many total doses are to be given, and how the patient is to be followed clinically.

UK usually is mixed in sterile saline. We will typically mix a standard vial of 250,000 U of UK in 100 to 250 ml of saline, giving a concentration of between 1000 to 2500 U/ml. We will generally divide this into three equal aliquots, although the volume is adjusted somewhat depending on the size of the collection to be treated. We administer these over the next 24 hours, refrigerating the unused doses and rewarming them to room temperature before instillation. The drainage catheter-typically 10 to 12-French in size-is flushed after instillation, clamped for about one hour, and then opened to closed system suction drainage (in most cases of empyema or hemothorax) or gravity drainage (for extrathoracic collections). If fluid output increases in response to fibrinolysis but drainage is incomplete, we repeat the steps listed above; we find it uncommon to treat with more than six total doses. We do not routinely follow hemoglobin/hematocrit or coagulation parameters, but may do so if the fluid output becomes bloody. Criteria for tube removal are no different from drainages without adjunctive lytic therapy and include successful evacuation of all fluid, obliteration of residual cavities, clinical improvement, and minimal amounts of drainage.

Reasonably similar protocols have been described by other authors. Moulton, for example, reported doses of 1000 or 2500 U/ml, volumes usually of 80 ml (but varying depending upon the size of the cavity), catheter clamping of 1 to 4 hours, and readministration immediately after unclamping or after 1 to 2 hours of suction drainage, with at least three separate administrations per day.18 Nakamoto and Haaga base their protocol for abdominal abscesses on abscess size, with a dose of 12,500 U for cavities up to 3 cm in diameter, 25,000 U for those of 3 to 5 cm in diameter, 50,000 U for a 5 to 10 cm diameter, and 100,000 U for those greater than 10 cm in diameter; they administer these doses every 8 hours for 4 days, and allow the tube to be clamped for 15 minutes after administration.1

What are the disadvantages of the use of fibrinolytic agents?

In addition to the potential risks of therapy discussed above, the cost of fibrinolytic therapy can be significant. This is particularly true of UK;2,18,35 at our institution, for example, a vial of 250,000 U of UK costs $343.18 (hospital cost), with cost to the patient of three to four times this amount. On the other hand, such cost could clearly be justified if the use of lytic agents results in a significant decrease in the length of hospital stay and/or improved drainage success without the need for additional therapies or tube manipulations.

Conclusion

While enthusiasm should be tempered by the realization that more studies-particularly studies with control groups-are needed, the use of fibrinolytic therapy in drainage of complicated fluid collections appears to be a safe and valuable addition to standard drainage techniques. AR

References

1. Nakamoto DA, Haaga JR: Percutaneous drainage of postoperative intra-abdominal ab-scesses and collections. In: Cope C (ed): Current Techniques in Interventional Radiology, pp 111-123. Philadelphia, Current Medicine, 1995.

2. Light RW: Parapneumonic effusions and empyema. In: Light RW (ed): Pleural Diseases, ed 3, pp 129-153. Baltimore, Williams & Wilkins, 1995.

3. Rotstein OD, Pruett TL, Simmons RL: Fibrin in peritonitis. V. Fibrin inhibits phagocytic killing of Escherichia coli by human polymorphonuclear leukocytes. Ann Surg 203:413-419, 1986.

4. Rotstein OD, Nathens AB: Peritonitis and intra-abdominal infections. In: Wilmore DW, Cheung LY, Harken AH, et al (eds): Scientific American Surgery, vol. 2, pp 1-24. New York, Scientific American, Inc., 1997.

5. Dunn DL, Simmons RL: Fibrin in peritonitis: III. The mechanism of bacterial trapping by polymerizing fibrin. Surgery 92:513-519, 1982.

6. Ahrenholz DH, Simmons RL: Fibrin in peritonitis: I. Beneficial and adverse effects of fibrin in experimental E. coli peritonitis. Surgery 88:41-47, 1980.

7. Hau T, Payne WD, Simmons RL: Fibrinolytic activity of the peritoneum during experimental peritonitis. Surg Gynecol Obstet 148:415-418, 1979.

8. Lahorra JM, Haaga JR, Stellato T, et al: Safety of intracavitary urokinase with percutaneous abscess drainage. AJR 160:171-174, 1993.

9. Sinclair SB, Rotstein OD, Levy GA: Disparate mechanisms of induction of procoagulant activity by live and inactivated bacteria. Infection Immun 58:1821-1827, 1990.

10. Jacques P, Mauro M, Safrit H, et al: CT features of intraabdominal abscesses. AJR 146:1041-1045, 1986.

11. vanSonnenberg E, Mueller PR, Ferruci JT Jr.: Percutaneous drainage of 250 abdominal abscesses and fluid collections. Part 1: Results, failures, and complications. Radiology 151:337-341, 1984.

12. Gerzof SG, Johnson WC, Robbins AH, Nabseth DC: Expanded criteria for percutaneous abscess drainage. Arch Surg 120:227-232, 1985.

13. Merriam MA, Cronan JJ, Dorfman GS, et al: Radiographically guided percutaneous catheter drainage of pleural fluid collections. AJR 151:1113-1116, 1988.

14. Moulton JS, Moore PT, Mencini RA: Treatment of loculated pleural effusions with transcatheter intracavitary urokinase. AJR 153:941-945, 1989.

15. Tillett WS, Sherry S, Read CT: The use of streptokinase-streptodornase in the treatment of chronic empyema. J Thorac Surg 21:325-341, 1951.

16. Park JK, Kraus FC, Haaga JR: Fluid flow during percutaneous drainage procedures: An in vitro study of the effects of fluid viscosity, catheter size, and adjunctive urokinase. AJR 160:165-169, 1993.

17. Coselli JS, Mattox KL, Beall AC: Reevaluation of early evacuation of clotted hemothorax. Am J Surg 148:786-790, 1984.

18. Moulton JS, Benkert RE, Weisiger KH, Chambers JA: Treatment of complicated pleural fluid collections with image-guided drainage and intracavitary urokinase. Chest 108:1252-1259, 1995.

19. Light RW: Malignant pleural effusions. In: Light RW (ed): Pleural