Abeloff's Clinical Oncology, 4th Edition

Part II – Problems Common to Cancer and its Therapy

Section F – Local Effects of Cancer and Its Metastasis

Chapter 60 – Effusions

Rosalyn A. Juergens,Alex I. Spira,
Julie R. Brahmer


Malignant Pleural Effusion




Common complication of cancer



Frequently seen with breast and lung cancers and lymphoma



Often asymptomatic




Presenting sign of malignancy in about one-half of patients who develop an effusion



Approximately 40 cases of malignant pleural effusion diagnosed for every 10,000 hospital admissions




Exudative effusions (high protein, lactate dehydrogenase, or cholesterol) are very concerning for malignancy.



Cytology is the gold standard.



If an exudative effusion exists but is undiagnosed by other means, video-assisted thorascopic surgery (VATS) should be used for evaluation of the pleura.




Cytology should be evaluated by an experienced pathologist.



Low pH (<7.3) is associated with burden of disease, less likelihood of response to pleurodesis, and decreased survival.




Asymptomatic effusions can be monitored.



Systemic chemotherapy can be effective in stable patients.



For patients symptomatic from their effusions, effective options include systemic therapy for patients with very responsive disease (e.g., hematologic malignancies, germ cell tumors, breast cancer, small cell lung cancer) or repeated thoracentesis.



For patients requiring frequent thoracentesis or who rapidly redevelop effusions, options include the following:



Pleurodesis with talc (favored), doxycycline, or bleomycin



Pleurodesis with placement of a drainage catheter only



VATS with mechanical abrasion



Open surgical procedure in highly selected patients



For patients refractory to pleurodesis, options include Denver-type drainage catheter or clinical trial

Malignant Pericardial Effusion




Less common complication of cancer than pleural effusion or ascites



Frequently seen with breast and lung cancers and lymphoma



Size of effusion not correlated with symptoms; rather, acuity of collection and compliance of pericardium most associated with tamponade physiology



Often a preterminal event




It is found in 2% to 30% of patients with cancer at autopsy.



Only about 10% to 20% of patients with pericardial metastases develop tamponade.




Cytology is the gold standard but not very sensitive.



Consider malignancy, comorbidities, and side effects of treatment (e.g., mediastinal radiation) in the differential diagnosis of a pericardial effusion in a cancer patient.




Cytology should be evaluated by an experienced pathologist.



Pericardial biopsy could be required if cytology is negative.



Patients should be evaluated for clinical signs of cardiac tamponade (hypotension, signs of low cardiac output, abnormal pulsus paradoxus).



Electrocardiogram should be evaluated for low voltages and electrical alternans.



Echocardiogram should be obtained for all patients.




Asymptomatic effusions can be monitored.



Systemic chemotherapy might be effective for stable patients, even those with solid tumors.



Those with hypotension/clinical instability should proceed to emergent pericardiocentesis and placement of pericardial drainage catheters.



Stable patients requiring drainage should undergo subxiphoid pericardiostomy.



Open procedures or pericardial sclerosis should be saved for refractory patients.

Malignant Ascites




Common complication of cancer



Frequently seen with ovarian cancer (most common cause), gastrointestinal malignancies, and carcinoma of unknown primary



Rarely life-threatening




Accounts for approximately 10% of all patients with ascites



Often (about 50% of the time) can be a presenting feature of malignancy




Ascites with low serum-to-ascites-albumin gradient is concerning for malignancy rather than portal hypertension.



Cytology is the gold standard.




Cytology should be evaluated by an experienced pathologist.



Patient should be assessed for symptoms and probable response to systemic chemotherapy.




Asymptomatic ascites can be monitored.



For patients symptomatic from their ascites, effective options include:



Diuresis with spironolactone/furosemide



Repeated paracentesis as needed



Patients with moderate long-term survival can be considered for either percutaneous tunneled drainage catheter or peritoneovenous shunt.


A malignant pleural effusion is a common complication of cancer and is encountered frequently by oncologists and other physicians caring for such patients. Their effects can range from causing no symptoms to being markedly symptomatic. Treatments must be tailored appropriately ( Fig. 60-1 ).


Figure 60-1  Treatment approach to malignant pleural effusions.



Collections of pleural fluid are likely to be directly related to tumor involvement of the pleura. For the patient with cancer, diagnostic accuracy in the assessment of pleural effusion is of the utmost importance. In many instances, the existence of a malignant pleural effusion drastically alters the treatment modalities that a patient will ultimately undergo; likewise, it drastically alters the likelihood of curing a patient with cancer. Nevertheless, many patients with cancer might have undergone surgery, have had radiation, or have other possible reasons for an effusion; there could therefore be alternative, benign reasons for an accumulation of pleural fluid. With this in mind, it is important to determine the appropriate cause of an effusion to provide optimal care.

Etiology and Pathogenesis

Pleural Space

The pleural space is a real space completely surrounding the lung up to the hilar root; the visceral pleura covers the lung and interlobar fissures, and the parietal pleura covers the chest wall, diaphragm, and mediastinum. The small amount of normal physiologic pleural fluid allows transmission of the breathing effort from the lung to the chest wall, thereby allowing respiration to occur. With the aid of a lavage technique in humans, each normal lung has been found to have approximately 0.13 mL pleural fluid/kg body mass (approximately 7 mL per lung). [1] [2]

Under normal conditions the body is highly efficient in controlling the amount of pleural fluid present; a 10-fold increase in the production rate of fluid results in only a 15% increase in pleural fluid volume.[3] Malignant effusions are caused by both increased entry and decreased exit of fluid. Increased fluid entry seems to be the principal abnormality, because there are multiple pathways and channels of fluid resorption. Among patients whose effusion reaccumulates rapidly after drainage, it is likely that the patient has mainly an increased production of fluid. Exit blocks probably account for a small number of patients (5% to 10%) with a malignant effusion and most often appear as a transudate upon analysis.[4] In evaluating the cause of an effusion, it is helpful to classify effusions according to their mechanism ( Table 60-1 ).

Table 60-1   -- Differential Diagnosis of Pleural Effusion by Mechanism



Metastatic malignancy

Intrathoracic malignancy (lung cancer)

Malignant mesothelioma


Bacterial, viral, fungal, mycobacterial, mycoplasmal, parasitic empyema

Pulmonary embolism

Connective tissue disorders

Drug induced (nitrofurantoin, procarbazine, amiodarone)

Esophageal rupture




Collagen vascular disorders

Rheumatoid arthritis, systemic lupus erythematosus, Wegener's granulomatosis, polyarteritis nodosa, Churg-Strauss syndrome

Dressler's syndrome (postmyocardial infarction syndrome)

Subdiaphragmatic abscess

Systemic capillary leak syndrome/adult respiratory distress syndrome

Vascular trauma

Postsurgical (including thoracic and abdominal surgeries)


Yellow-nail syndrome



Congestive heart failure

Superior vena cava syndrome

Massive pulmonary embolism



Metastatic malignancy




Post–lymph node irradiation

Meig's syndrome

Peritoneal dialysis


Hypoalbuminemia (often from cancer cachexia)



Nephrotic syndrome



Trapped lung (infection, malignancy)




The molecular biology of pleural effusions has begun to be understood, with vascular endothelial growth factor (VEGF) emerging as a major role player.[5] Because it induces endothelial vasodilation and enhances the permeability of the mesothelium 50,000 times more potently than histamine, VEGF is thought to be a major, if not the most important, cytokine in the etiology of effusions.[6] It might one day be part of the diagnostic evaluation of effusions. Novel approaches that use anti-VEGF mechanisms will probably be evaluated once this is better understood.[7]


More than a million pleural effusions occur each year, with about one-half arising from congestive heart failure (CHF). About one-fifth of effusions are malignant, but malignancy is the most common cause of symptomatic effusions. Under-reporting in most series might also be a significant issue. The causes of malignant pleural effusions are shown in Table 60-2 and reflect the incidence of cancers in general, with lung and breast cancers at the top of the list. [8] [9] [10] [11] [12] [13] [14]

Table 60-2   -- Malignant Neoplasms Associated with Pleural Effusion




Total malignant effusions



Lung cancer



Breast cancer



Lymphomas and leukemia



Unknown primary (adenocarcinoma)



Unknown primary (all types)



Gastrointestinal tract



Reproductive tract



Genitourinary tract



All other[*]




Includes causes of malignant effusion (each less than 1%): endocrine, head and neck cancer, mesothelioma, soft-tissue sarcoma, bone cancer, and myeloma.


Diagnosis and Evaluation of a Pleural Effusion

Pleural effusions are apparent on a chest radiograph when a volume of 200 mL or greater of fluid is present.[15] Oncologists’ patients can be generally subdivided into those with clear advanced disease and probable malignant pleural effusion and those who have a history of malignancy but in whom the effusions may or may not be malignant. Among patients with advanced cancer, there is often no question as to whether the effusion is directly related to the malignancy, or the question is unimportant in terms of the patient's overall clinical management. In such a case, one can simply observe asymptomatic patients. For the patient in which an effusion could be malignant, however, accurate diagnosis becomes imperative for patient care and management, and one must consider all alternative causes in the etiology of the effusion. These include the possibility of other coexisting illnesses (e.g., CHF, infection), consequences of treating a prior malignancy (e.g., sequelae of prior surgery, radiation, or chemotherapy), and the possibility that the effusion is indeed malignant. Most locally advanced solid-tumor malignancies are considered incurable if the effusion is malignant, and treatments and goals of treatment will thus become dramatically different. In some cases, however, such as for childhood tumors (e.g., Ewing's sarcoma) or germ cell tumors, patients might remain curable but treatment plans can become radically different. Hematologic malignancies (e.g., lymphomas) might be upstaged, and this step might or might not affect the planned treatments. Lung cancer, the most common etiology of malignant effusions, can cause effusions both via direct involvement of pleura by tumor with increased fluid production and by decreased exit via lymphatic blockage. Patients with non-small-cell lung cancer and transudative effusions on multiple evaluations of their pleural fluid are, in fact, considered early stage and are usually treated with curative intent, as if the effusion were caused by lymphatic blockage and not direct involvement of the pleura. Those with exudative effusions are considered to have advanced disease, even if malignant cells cannot be found, and are treated in a palliative manner. Patients with small cell lung cancer with exudative effusions are considered to be “extensive stage” because the field cannot be covered in a single, tolerable radiation port; these patients are treated in a palliative manner with chemotherapy alone. Patients with other nonmalignant reasons for effusions (e.g., large amount of ascites and cirrhosis or CHF) may be observed, but the physician should have a low threshold for reevaluation should clinical conditions change (e.g., a new fever or new shortness of breath) or if the effusion fails to resolve with appropriate measures.[16]

History and Physical Examination

Patients with pleural effusions usually present with dyspnea, chest pain, cough, or orthopnea. Other symptoms, such as hemoptysis, fever, and dysphagia, might also occur but are less common. Physical examination usually demonstrates dullness to percussion, decreased or absent breath sounds, and absence of fremitus. As a result of atelectasis of the adjacent lung, one can often hear crackles at the superior borders of the effusion. Both sides should be listened to equally to determine whether the effusion is unilateral or bilateral. Bilateral, symmetric effusions might be more associated with volume overload states such as CHF, although this is a nonspecific finding. The examination should also be alert for signs of systemic infection or other clues about the etiology of the effusion.

Radiologic Evaluation

The use of a chest radiograph to evaluate a suspected pleural effusion is the most valuable and easiest tool. Lateral views should be obtained along with posterioanterior views, because the latter miss as much as 500 mL of fluid hidden behind the dome of the diaphragm. At least 75 mL of fluid is needed to obscure the posterior costophrenic angle; up to 175 mL of fluid is required to obscure the lateral costophrenic angle. If the effusion reaches the fourth rib, 1000 mL is present; as little as 10 mL can be seen on a decubitus film. In terms of sampling for thoracentesis, a general rule of thumb is that an effusion that is thicker than 1 cm on a decubitus film is large enough for sampling by thoracentesis.[17] Computed tomography (CT) scanning is even more accurate in detecting small effusions, including as little as 2 mL of fluid. The volume of the fluid present can best be determined radiographically by using three-dimensional reconstruction.[18] Neither magnetic resonance imaging (MRI) nor CT scan can distinguish transudates from exudates accurately, although both can be helpful in evaluating the pleural contents for masses, nodules, and pleural-based thickening once the fluid is removed.[19] Ultrasound is useful in evaluating for the presence of an effusion and as a guide during thoracentesis. Ultrasound also may aid in distinguishing an exudate (echogenic) from a transudate (anechoic), although this finding is not definitive.[20]


Among patients with an effusion that requires evaluation, the radiograph should be evaluated to determine whether the effusion is loculated. If the effusion is not free-flowing on the decubitus film and is therefore loculated or if the volume is insufficient to attempt a safe, unguided thoracentesis, ultrasound guidance should be used. [21] [22] When obtaining fluid, samples should be collected into tubes containing heparin to prevent clot formation; cloudy fluids should be centrifuged before the analysis described shortly.[23] Relative contraindications to thoracentesis include a skin infection at the site of needle insertion, an abnormal coagulation profile (prothrombin time or partial thromboplastin time >1.8 times normal, platelets <25,000/mm3, or creatinine >6 mg/dL), or uneasiness of the operator in performing the procedure.[24] Mechanical ventilation does not seem to increase the risk of pneumothorax, but there could be a higher likelihood of developing a tension physiology if a pneumothorax does occur.[25] Complications of the procedure include pain, bleeding (hemothorax or hematoma), infection, splenic or hepatic laceration, seeding of the needle tract with tumor, or, most commonly, pneumothorax.[26] Studies show that in thoracenteses performed by trained physicians, pneumothoraxes occur approximately 10% of the time, with a much higher rate when performed by trainees. [16] [25]Small pneumothoraxes can be observed in nonventilated patients, whereas larger ones might require chest tube decompression. Although chest radiograph is often obtained after thoracentesis, it is not required for patients who are clinically asymptomatic after the procedure (absence of cough, chest pain, shortness of breath, and a normal lung examination), who do not experience aspiration of air, and for whom only one needle pass was used during the procedure.[27] The use of a therapeutic thoracentesis is discussed shortly.

Evaluation of Fluid: Transudates and Exudates

Fluids are classified as either transudates or exudates. Transudates arise from a disturbance in the hydrostatic or colloid forces across a barrier and have low protein and cellular contents. By definition this is a noninflammatory process without pleural disease. A classic example of a transudative effusion is one due to CHF. Transudates may be thought of as volume overload states and are in general benign. Exudates, on the other hand, are formed by active secretion, inflammation, leakage, or permeability of a lymphovascular barrier and have a high cellular or protein content. They have a broader list of potential causes, can be more serious in their effects, and mandate a more thorough evaluation. In reality, most effusions that will be referred to an oncologist will be exudative in nature, because transudative effusions often present with a known systemic disease—for example, heart failure, cirrhosis, or nephrotic syndrome.

Dividing effusions into one of these two categories is conceptually useful in classifying patients with unknown effusions. In 1972 Richard Light published what has become the classic criteria for the biochemical classification of a pleural effusion and now is commonly known as Light's criteria.[28] According to this system, the values of lactate dehydrogenase (LDH) and protein are measured both in the pleural fluid and in serum, and the effusion is classified as an exudate if any one or more of the following three criteria are met:



Ratio of pleural protein to serum protein greater than 0.5



Pleural LDH greater than two-thirds of the upper limit of normal for the serum reference range



Ratio of pleural LDH to serum LDH greater than 0.6

Criterion number 2 reflects a modification from the original published manuscript to account for the wide variation in the normal range of LDH in various reference laboratories. [23] [29] A recent reevaluation of Light's criteria comes from Keller, demonstrating that Light's criteria have an overall sensitivity for an exudate of near 100%, but a specificity of only approximately 80%.[30]


The level of cholesterol is higher in exudative than transudative effusions for unclear reasons. Costa and colleagues were able to demonstrate that a cholesterol concentration greater than 45 mg/dL and an LDH greater than 200 IU/L in the pleural fluid alone identified exudates with a very high sensitivity (99%) and specificity (98%) without the requirement for a serum sample.[29] Using a pleural cholesterol to serum cholesterol ratio threshold of 0.3, Valdes and associates were able to provide fewer misclassifications than when applying Light's criteria.[31] In short, the level of cholesterol in pleural fluid seems useful in distinguishing exudates from transudates.

Optimal Testing

Light's criteria have stood the test of time and are the most widely regarded. A meta-analysis of eight studies including 1448 patients was unable to define an optimal diagnostic technique.[32] Based on these results, a 21st-century version of Light's criteria, modified only slightly from Light's original work by the addition of cholesterol as a marker, could be suggested as the best means of determining an exudate. According to these updated criteria, a patient can be said to have an exudate if any of the following criteria are met[32]:



Pleural fluid protein greater than 2.9 g/dL



Pleural fluid cholesterol greater than 45 mg/dL



Ratio of pleural fluid LDH to serum LDH greater than 0.6

The aforementioned meta-analysis demonstrated no superior combination (paired or triplet) to any one of these single tests. In most cases, particularly in the presence of obvious, known cancer, one should not have to delve beyond these criteria in the initial evaluation of a pleural effusion.

Other Testing

There are other situations in which a specific test might not differentiate between an exudate and a transudate but still might be helpful in differential diagnosis. However, although these tests are useful, they are not 100% specific but merely provide the clinician with some general clues that must be evaluated further.


The first step in the evaluation of pleural fluid is visual observation as it is removed. Strawlike fluid is more likely to be benign and transudative in nature. Fluid containing puslike material is more likely to be infective in nature. Bloody fluid is related to causes such as pulmonary infarction, malignancy, or a postoperative state. In general these clues can be helpful, but no diagnosis can be made definitively.


In a diagnostic thoracentesis, pleural fluid pH should always be measured. For proper analysis, pleural fluid should be collected into a single syringe and a sample then transferred to a smaller, heparinized syringe for analysis.[33] Analysis should be via a blood gas machine, not on litmus paper, because the latter is unreliable and not an acceptable alternative.[34] Normal pleural fluid pH ranges from 7.60 to 7.64. In one study, 46 patients with a pleural fluid pH less than 7.30 all had an exudate and one of the following six diagnoses: malignancy, empyema, collagen vascular disease, tuberculosis, esophageal rupture, or hemothorax.[35] Furthermore, patients with known malignant effusions and a pH less than 7.30 have a worse prognosis, shorter mean survival, poorer response to tetracycline pleurodesis, and a high rate of finding malignant cells on initial fluid cytology.[36] It should be remembered that patients with a parapneumonic effusion with either a pH less than 7.3 or gross pus should be considered for chest tube placement for drainage. The reader is referred to other texts for the management of empyemas.


Cell counts are often obtained on pleural fluids but are rarely useful or diagnostic. Very high nucleated cell counts (>50,000 per milliliter) are associated with complicated parapneumonic effusions and empyema but rarely are diagnostic.[16] Because most cells are normally neutrophils or monocytes, a predominance of lymphocytes (>50%) should make one more seriously entertain the idea of a carcinomatous pleural effusion, and greater than 85% lymphocytes should make one entertain the diagnosis of lymphoma, sarcoidosis, chylothorax, rheumatoid pleurisy, or yellow nail syndrome. [16] [37] An increase in pleural fluid eosinophilia (>10% of nucleated cells) might be associated with benign disease (hemo- or pneumothorax), but also can be associated with all types of malignancy.[38] The presence of mesothelial cells is not helpful in terms of diagnosis. [37] [39]

Pleural fluid cytology is the simplest and most definitive method of diagnosing a malignant effusion. The sensitivity depends on the type of malignancy, extent of disease, and experience of the cytopathologist. Fluids should be concentrated first for optimal detection of malignancy.[40] In general, the sensitivity is on the order of 62% to 90% and, as the gold standard, pleural fluid cytology is virtually 100% specific in the hands of an expert cytopathologist.[41] If an effusion demonstrates carcinoma and breast cancer is a diagnostic possibility, the cytologic specimen can be stained for estrogen and progesterone receptors as a means of both diagnosis and selection of potential treatment. In short, the presence of an abnormal cell population should prompt a further workup for the aforementioned causes, including malignancy.

Other tests have been evaluated in numerous studies as a means to refine the diagnosis of pleural effusions. For example, an elevated adenosine deaminase concentration has a high association with tuberculosis.[42] Elevated lipids (triglycerides) can be associated with a chylothorax and obstruction of the thoracic duct by any number of means, including malignancies.[43] In clinical experience, this and other tests (e.g., creatine kinase, LDH isoenzyme analysis, β2-microglobulin, albumin, ferritin, lysozyme, and others) are rarely used, are rarely associated with malignancy, and are therefore discussed elsewhere.[23] Other tests including tumor markers (carcinoembryonic antigen [CEA], CA 19-9) have also been evaluated for use in determining the etiology of effusions but have not been found to be sensitive or specific.

Evaluation of a Suspected Malignant Effusion

Although most malignant effusions occur among patients with known cancers, they can be the first indication of the presence of malignancy in as many as 30% of patients.[44] In some patients an effusion often can be the only site of a potential malignancy after a thorough evaluation. Once a patient has been diagnosed with an exudative effusion, a malignant cause must be high on the list of differential diagnoses. A thorough history and complete physical examination must be performed, with careful attention to any potential causes or risk factors of malignancy. Once this evaluation takes place, the physician must gather some definitive evidence to institute appropriate evaluation. Frequently, such evidence involves consultation between the patient's primary physician and either a pulmonologist or an oncologist. Many clues about the etiology of the effusion are obtained with the performance of further evaluations, such as chest radiographs, CT scans, or mammograms. Alternatively, cytologic evaluation of an exudative effusion can reveal the presence and type of malignancy directly.

It is important to keep in mind the most common malignancies that cause effusions. Not surprisingly, a majority of these are caused by lung cancer. Breast cancers and lymphomas also cause a significant number of these, with approximately one-third of all malignant pleural effusions being caused by other types of malignancies (see Table 60-2 ).[45] Appropriate evaluations should be performed as indicated (e.g., careful breast examination and mammogram in women with effusions; particular attention for lung cancer among patients with a history of smoking). It should be noted that there is a recently established entity, primary effusion lymphoma. This is found in patients positive for the human immunodeficiency virus; it is associated with human herpesvirus 8 infection and has a pathogenesis similar to Kaposi's sarcoma.[46] This syndrome also can include pericardial effusions and ascites. Patients with mesothelioma often have a history of asbestos exposure and show evidence of both effusion and pleural thickening on CT scan. These patients can be considered for CT-guided biopsy of the appropriate areas, which is often effective in making a diagnosis.[47] A small percentage will be diagnosed with carcinoma of unknown primary, whose management, usually via chemotherapy, is also discussed in Chapter 98 . Even after an extensive evaluation, results can be nondiagnostic, and the patient must undergo further evaluation to determine the etiology of the effusion because no other cause has been identified although malignancy is still suspected. These further evaluations will now be discussed.

Closed Pleural Biopsy

These refer to blind, percutaneous biopsies of the parietal pleura using a special needle, such as a Cope's needle. Unfortunately, among patients with a cytology-negative malignant effusion, the yield of this procedure is only about 7%.[48] Novel approaches to this procedure, such as the use of brushings or a Tru-cut needle, might ultimately prove to be better diagnostically, but the procedure is currently rarely used in such situations because of its poor yield. [49] [50]


Video-assisted thoracoscopic surgery (VATS) is the most commonly used procedure for thoracoscopy in the United States and is usually performed by thoracic surgeons. This procedure, using several ports and trochars, requires general anesthesia with single-lung ventilation and many single-use disposable instruments.[51] Although VATS is well tolerated, it does have some risks and carries with it a significant expense, due to general anesthesia and the requirement for expensive instruments. On occasion, it requires conversion to an open procedure if there are significant adhesions or if there are undue risks noted with the insertion of a thoracoscope.[45] Single-lung ventilation is also required and might be difficult for patients with significant lung disease.

With these techniques, fewer than 10% of malignant pleural effusions go undiagnosed, and thus thoracoscopy of some sort has become part of the evaluation algorithm when necessary if other methods fail.[52] Interestingly, a recent technique uses a semi-rigid pleuroscope, which is similar to bronchoscopes currently in use by pulmonologists but easier to use. These can be used to drain and pleurodese effusions and might ultimately decrease the need for VATS in diagnosing potentially malignant effusions if comparison studies demonstrate its effectiveness.[53]


For most malignancies, the existence of a malignant effusion places the patient into a noncurable, advanced staging category, but one that is often treatable nonetheless. In light of this a palliative approach is often the mainstay of therapy, with several important exceptions that will be addressed in the discussion that follows. For patients with relatively small effusions that do not cause a high degree of dyspnea or impairment of functional status, consideration of systemic therapy of the malignancy, usually chemotherapy, is indicated. If, on the other hand, dyspnea is a primary concern, immediate management of the effusion is necessary. Unfortunately, for many patients with advanced disease, chemotherapy might not be able to provide rapid enough resolution of dyspnea or other symptoms, and thus one must use mechanical means of reducing the effusion. First and foremost, the clinician should consider the individual patient's situation, with particular attention to overall prognosis, prior therapies (if any), age, and performance status. The physician must also weigh the likelihood of progression of other systemic disease during the time required for resolution of an effusion. For patients with large effusions who are in the terminal stage of their disease, supplemental oxygen with hospice referral might be the only appropriate intervention.

Systemic Therapy in Specific Malignancies

Many advanced solid tumors do not respond well to chemotherapy, so primary management of the effusion is often required. There are notable exceptions. Small cell lung cancer is unlikely to be curable, but effusions from this cause respond well to chemotherapy. [54] [55] Testicular and other germ cell neoplasms are not only responsive but also very curable, even at an advanced state, and thus early systemic treatment should be used if at all possible.[56] Even patients with advanced germ cell tumors have good responses and should be considered for aggressive systemic therapy. Breast cancer might respond well to hormonal therapy or chemotherapy, particularly among patients who have not received prior systemic therapy. Mesothelioma is a unique tumor that arises directly from pleural or peritoneal surfaces. Effusions are often associated with the tumor and are not included in the staging system. Patients amenable for aggressive local therapy, even with an effusion, should be considered for such therapy. Effusions from hematologic malignancies, such as lymphoma, are often highly responsive to systemic chemotherapy. Treatment of effusions related to hematologic malignancies are generally focused on treatment of the disease as a whole rather than treatment directed at the effusion directly. Because patients with solid-organ tumors and advanced disease have relatively poor responses to chemotherapy overall, early direct management of the effusion among patients with dyspnea should be considered at the outset for patients who are unlikely to achieve rapid responses to chemotherapy.


It should be remembered that the pharmacokinetics of drugs in the presence of effusions are poorly understood, because drug might accumulate in effusions and only slowly redistribute throughout the body. Methotrexate is a classic example of this phenomenon. Methotrexate is slowly released from all of the “third spaces,” and those of large volumes (ascites, pleural effusions, or anasarca) might dramatically prolong the terminal half-life and lead to potential increased toxicity.[57] There are no direct guidelines for the use of methotrexate or other chemotherapies when large effusions exist, but the clinician should consider drainage of the effusion before the use of methotrexate, or the use of alternative therapies. Patients with pleural effusions receiving new chemotherapy agents should be monitored closely for treatment-induced toxicity.

Effusion-Targeted Therapy


Therapeutic thoracentesis might serve as the main or sole therapy for management of an effusion in many patients. Thoracentesis, unlike other procedures, might permit rapid relief of symptoms without need of hospitalization. If systemic disease is a significant concern, thoracentesis might allow for a window of opportunity in which to gain control over a patient's symptoms before systemic chemotherapy is given. Similarly, it might also allow for palliation of symptoms for those patients with far advanced disease.

Therapeutic thoracenteses are performed similarly to diagnostic procedures. Although a pneumothorax usually occurs when air leaks into the pleural space through the needle, it can also occur if the visceral pleura is lacerated with the needle. The latter complication may be avoided if a plastic catheter is threaded through the needle, directed inferiorly, and the needle is then removed; this procedure minimizes the chances that a pneumothorax will occur via perforation when the visceral and parietal surfaces become opposed after fluid removal. If a pneumothorax occurs the leak usually seals itself quickly, but closure can be hastened by having the patient lie on the affected side, which decreases the pressure gradient between the alveoli and the pleural space.[58] The volume of pleural fluid that can be removed safely is unknown. One study by Light and coworkers demonstrated that pleural fluid removal could be considered safe as long as pleural fluid pressure does not decrease below -20 cm H2O.[28] Because most clinicians do not measure pleural pressure, current recommendations are that no more than 1.0 to 1.5 L of fluid be removed at any one time, and that amount only if there are no signs of any adverse events such as dyspnea, pleural pain, or cough, which are indications of the rapid restoration of a strong negative pleural pressure. Patients with contralateral mediastinal shift can have larger amounts of fluid removed safely, because these adverse events are unlikely to occur. Rapid reexpansion might bring on the phenomenon of reexpansion pulmonary edema, which is due to the rapid restoration of capillary permeability through unknown mechanisms; this can occur when air or fluid is removed from the pleural space.[58] Patients with ipsilateral mediastinal shift are unlikely to obtain significant relief from thoracentesis, because this condition indicates a large amount of either trapped lung or mainstem bronchial occlusion.[46]

There have been no randomized or comparison trials examining the use of repeated thoracentesis compared with the use of other procedures. It is also difficult to predict the length of time within which an effusion might recur; some patients have rapid recurrence (within days), whereas others might have recurrence over a period of weeks. Most oncologists and pulmonologists often make an initial attempt at a therapeutic thoracentesis to allow immediate symptomatic relief and time for systemic therapy to take effect, as well as to gauge the extent to which a pleural effusion causes dyspnea. For patients whose effusions recur rapidly, more invasive procedures might be required; those whose effusions recur more slowly might be managed solely with repeat thoracentesis.


In general, radiation of the hemithorax, as a means of controlling an effusion, is contraindicated in most patients with malignant pleural effusions. This is because of the high incidence of radiation pneumonitis that is likely when a sufficient dose of radiation is given to large areas potentially involved with malignant effusions.[59] Certain situations, such as lymphatic obstruction from focal areas of lymphadenopathy (which may arise in lymphoma or in lung cancer), might benefit from radiation applied to these specific areas. Radiation doses are dependent on the nature of the malignancy.


Pleurodesis is intended to achieve a “symphysis between the parietal and visceral pleura, in order to prevent accumulation of either air (pneumothorax) or fluid (pleural effusion) in the pleural space,”[60]and malignant pleural effusions are the largest indications for pleurodesis. [60] [61] The mechanism by which this symphysis occurs is poorly understood, but in general it depends on pleural irritation to cause a cycle of inflammation, activation of the pleural coagulation cascade, fibrin placement and fibroblast recruitment, and, finally, collagen deposition that ultimately results in the fusion of both pleural surfaces. [60] [62] [63] Nevertheless, the exact mechanisms and factors that influence pleural sclerosis and effect pleurodesis are not well known and require further research. Many agents have been used in the past with various success rates reported; much of this information is based on personal and anecdotal experiences. These agents include, as a minimum list






























Nitrogen mustard



50% glucose and water












Radioactive colloidal gold



Autologous blood



Fibrin glue



Bacille Calmette-Guérin



Silver nitrate



Killed Corynebacterium parvum[60]

The exact mechanism by which these agents effect pleurodesis probably differs slightly from one agent to another, but all lead to a final common pathway that activates the pleural coagulation cascade and the appearance of a fibrin network that yields a symphysis between the two surfaces. [61] [62] [63] The various agents have different properties and methods of administration.


It should be recalled that patients must demonstrate that symptoms (usually dyspnea) respond to drainage of the pleural fluid, because patients with malignancy often have numerous reasons for dyspnea—pulmonary metastases, anemia, pulmonary embolus, trapped lung, or poor gas exchange—and might be unlikely to benefit from a resolution of the effusion. Thus, it is important for patients to undergo a trial of therapeutic thoracentesis initially rather than proceed directly to chest tube drainage. At the time that pleurodesis is performed, the effusion must be drained. For successful sclerosis, there should be evidence of complete lung reexpansion after initial drainage. Traditionally one gauged the optimal time for sclerosis when there was minimal pleural fluid drainage (<150 mL per day), but this seems to be less relevant for success in sclerosis than does the confirmation of complete radiologic lung reexpansion.[64]

Traditionally chest tube drainage has been performed with a standard-size (24–32 Fr) chest tube. Tubes in this size range are associated with a great deal of pain and discomfort. More recent work has supported the use of smaller bore (8–16 Fr) catheters. A prospective, randomized study of 18 patients found improved comfort and no difference in success or complication rates when small-bore catheters were used.[65] Small-bore tube placement is an emerging standard of care.[66]

Sclerosing Agents


Tetracycline was the most widely used sclerosing agent until it was discontinued by the manufacturer in 1992, although it might still be available in some countries. [67] [68] Doxycycline has been recommended as a replacement, but there are no direct studies comparing the two agents. Historical comparisons demonstrate similar success rates with doxycycline compared with tetracycline, both on the order of 80% to 85%. [69] [70] Most studies and investigators have used 500 mg of doxycycline mixed in 50 to 100 mL of sterile saline, and the primary complication is pain related to the doxycycline, requiring either narcotics or conscious sedation. [45] [69] [70]


Talc is currently the most widely used, the best studied, and probably the most controversial sclerosing agent. A major advantage of talc is the cost; the average wholesale price for the amount of talc typically used for pleurodesis procedures is less than $1.00. Sterilization of the talc raises the cost to between $5 and $20, dependent on the methods used, and it remains sterile for at least a year on pharmacy shelves.[71] The cost of talc is therefore significantly less than for any of the other agents used. Talc may be introduced via slurry (i.e., mixed with saline) in the chest tube used to drain the effusion, or insufflated (poudrage) with a bulb syringe or atomizer at the time of a thoracoscopic drainage procedure. (Thoracoscopy will be discussed shortly.)

In 1994 Kennedy and Sahn reviewed all the published series using talc as a pleurodesis agent and found a 91% success rate (659 of 723 patients) for pleurodesis, with success judged by various clinical and radiologic findings.[72] Doses of talc ranged widely, from 1 to 14 g per procedure. There was no difference between poudrage and slurry. Animal studies suggest that concomitant use of steroids can decrease the efficacy of talc pleurodesis, and a small study in humans demonstrated a small, not statistically significant decreased response to talc slurry when steroids were used. [73] [74] Although these studies are not conclusive, steroids should be discontinued or the dose reduced as much as possible before pleurodesis procedures when talc or any pleurodesis agent is used.

Adverse effects from talc are variable. Pain ranging from nonexistent to severe is not uncommon; there is an overall reported incidence of 7%.[75] Fever, up to 102°F, occurs within the first 12 hours and can last up to 72 hours. [16] [75] Empyema is an occasional complication, more often with talc slurry than with poudrage. Cardiovascular complications (e.g., arrhythmia, chest pain) have been noted, but it is difficult to assess whether these are a result of the talc pleurodesis procedure itself or the patients’ comorbidities. [72] [74]

Pulmonary complications constitute the largest and most significant group of complications associated with talc pleurodesis. Acute respiratory distress syndrome (ARDS), pneumonitis, and respiratory failure all have been reported after pleurodesis with talc. [75] [76] The mechanisms by which talc might produce acute lung injury are unclear. One hypothesis is that the pneumonitis is related to the systemic absorption of talc; this is supported by the findings of Rinaldo and colleagues[77] and others who were able to demonstrate talc particles both in bronchoalveolar lavage fluid after talc pleurodesis and in virtually every organ at autopsy of one patient. [76] [77] [78] [79] One of the most concerning studies comes from Rehse and associates,[80] who found that in a series of 78 patients, 33% developed respiratory complications or death after talc pleurodesis, and 9% developed ARDS rates that are significantly higher than those previously reported. Some investigators feel that the incidence of pulmonary complications is related to the dose of talc used for pleurodesis. Talc pleurodesis is still commonly used, however, and the relative merits of this in comparison to other methods are fiercely debated among experts and will be discussed later in this chapter. [78] [81]


Bleomycin has been studied since the 1970s for its use as a sclerosing agent. [82] [83] The recommended dose for intrapleural administation is 60 U or 1 U/kg body weight, and is often reduced to 40 U/kg in the elderly.[84] Administration and its side effects are similar to those with talc pleurodesis, with pleuritic pain, rigors, fever, and mild nausea as the main concerns.[84] Although it also is used commonly as a chemotherapeutic agent, only 40% is systemically absorbed from the pleural cavity, and bleomycin is not myelosuppressive when used in pleurodesis.[82] Several studies report mixed results for bleomycin compared with other agents. A large meta-analysis (1168 patients) from Walker-Renard and coworkers[85] demonstrated a 54% success rate with bleomycin versus a 67% success rate with tetracycline and its derivatives (doxycycline, minocycline) and a 93% success rate with talc. Zimmer and colleagues,[86] in a subsequent prospective randomized trial, did not demonstrate a statistically significant difference in efficacy between talc (90% successful) and bleomycin (79% successful, P = 0.388). Patz and associates[84] directly compared bleomycin and doxycycline with a small-bore catheter and likewise did not find any statistically significant difference between the two agents. It is important to note that there are large cost differences between the different agents. In the aforementioned study by Zimmer and coworkers, the cost of the bleomycin was $955 per patient, compared with only $12 per patient for talc.[86]


Interestingly, two studies demonstrated that one could obtain relatively effective pleurodesis with thoracoscopy and chest tube placement for several days without instillation of a pleurodesis agent; these studies demonstrate a combined 62% success rate among 61 total patients. [87] [88] Talc insufflation at the time of thoracoscopy does indeed increase the success rate to more than 90% of patients, however.[72] The fact that one can achieve decent pleurodesis and modest success rates without the use of any pleurodesis agent highlights our relatively poor understanding of the mechanism of pleurodesis and suggests that irritation of any sort can in fact lead to sclerosis and symphysis of the two pleural membranes. One potentially superior alternative to simple chest tube placement is to perform mechanical abrasion of the pleura. There are, however, no large series evaluating mechanical abrasion at the time of thoracoscopy for effectiveness, nor any comparing mechanical abrasion with any other techniques.[78] [79] Mechanical abrasion can be performed with a thoracoscope (usually VATS) or via an open procedure. Pleural abrasion, in conjunction with bleb stapling, has been shown to be effective in preventing recurrent pneumothorax as a method of sclerosis in this nonmalignant entity.[89]


It has been stated that low pleural fluid pH can identify patients who might experience a low likelihood of success with pleurodesis.[61] Low pH is thought to be a marker of increased metabolic activity of intrapleural tumor, and as such it represents a larger tumor burden.[36] Large tumor burdens might prevent apposition of the pleural membranes. Heffner and colleagues[90] reanalyzed individual patient data from both published and unpublished studies in 2000. Their analysis of 433 patients demonstrated that pleural fluid pH was the only independent predictor of pleurodesis failure, with an odds ratio of 4.46 when pleural fluid pH was lower than 7.28. A pH value of 7.15 or lower had a positive predictive value of 45.7% for pleurodesis failure. Although pH certainly helps predict pleurodesis failure, it is only of modest benefit in predicting outcome in individual patients, in that even those patients with a low pH in the analyzed studies had a greater than 50% success rate with pleurodesis. Because of these limitations, fluid pH should only be considered as an additional piece of information, not as a means to totally rule out pleurodesis.

Surgical Options Including Shunts

There are several surgical procedures—parietal pleurectomy, decortication, and pleuropneumonectomy—that may be attempted for management of a malignant pleural effusion. These procedures carry with them major morbidities, and they have proven to be no better than pleurodesis alone.[91] Surgical palliation might, however, be an option for some patients, particularly those who fail chemical pleurodesis, patients with loculated effusions, or those with large tumor rinds on the pleural surface. VATS permits direct visualization of the entire pleural surface, the potential for mechanical abrasion of the pleural surface, and removal of some adhesions and loculations. Talc poudrage via insufflation may be performed, and a pleuroperitoneal shunt may be inserted as well.[92] Surgical intervention should also be considered if lung reexpansion does not occur promptly after removal of an effusion via chest tube drainage; failure to reexpand suggests a rind of malignant tissue surrounding the lung, which may be further confirmed by a radiograph demonstrating a large effusion without mediastinal shift.[91] If such a cortex is seen at the time of VATS, it may be removed by conversion of the VATS to an open thoracotomy. This procedure might make pleurodesis and the restoration of a trapped lung possible. There is, however, a perioperative mortality of 12% associated with open procedures for decortication in these situations; therefore, this choice of procedure must be limited to an appropriate patient population.[91] Pleuroperitoneal shunts have a complication rate of approximately 12%, which is manifested mainly by shunt occlusion requiring replacement.[93] Peritoneal seeding of intrathoracic malignancy is a potential hazard but has not been definitively studied.[91]

Intrapleural Therapy

The rationale of intrapleural therapy is very appealing: Agents given directly into the pleural space can be directly toxic to tumor cells or are stimulants of the immune system that would ultimately be cytotoxic to tumor cells, thereby causing resolution of an associated malignant effusion. It should be noted that it is difficult to separate the cytotoxic effects of intrapleural therapy from pleural irritation to determine whether resolution of an effusion is due to cytotoxicity or pleurodesis. Bleomycin is a classic example of this phenomenon in that it is used as a sclerotic agent but also posseses antineoplastic activity. Intrapleural therapy, particularly chemotherapy, offers hope for patients who have a malignancy confined to the pleural space (e.g., mesothelioma) or for patients who have a higher disease burden in this location compared with other sites. A similar approach of locally directed chemotherapy has been used with particular success in the abdominal cavity, in cases of intraperitoneal chemotherapy for ovarian cancer.

Most agents that have been tried over the years have had limited success, and thus the idea of intrapleural therapy is still highly experimental. In terms of cytotoxic chemotherapy, many agents have been tried, including intrapleural 5-fluorouracil (5-FU), mitomycin, cisplatin, cytarabine, doxorubicin, and etoposide.[85] Shoji and associates[94] recently reported their phase II results using repeated intrapleural chemotherapy with an implantable access system (an infuse-a-port, similar to that used for subcutaneous venous access), placed by a VATS procedure. Patients received biweekly 5-FU and cisplatin. Such a system was used previously for the administration of interferon-γ in the treatment of malignant mesothelioma by Driesen and coworkers[95] and allows for repeated, easy administration of intrapleural drugs with minimal catheter-related toxicity. There was virtually no systemic toxicity, not even of the type that is usually associated with systemic therapy of these drugs, although the doses (5-FU 250 mg, cisplatin 10 mg) were far less than those typically used intravenously. Because of this observation, such a system warrants further testing as a method of both pleurodesis and treatment (neither of which was an endpoint in this dose-finding trial), particularly for tumors sensitive to agents that can be given intrapleurally.

As described in virtually all of these reports, the success rates with these cytotoxic agents are less than those seen with talc, doxycycline, and even intrapleural bleomycin. There is the potential for systemic absorption, and consequently potential associated systemic toxicity from such agents. The costs of antineoplastic agents are much higher than those for doxycycline and talc. Because talc is well studied, inexpensive, effective, and readily available, it would be difficult to find a superior agent for use solely in the treatment of effusions; for these reasons, intrapleural therapy still remains experimental excepting the use of agents that cause pleurodesis.

Long-Term Drainage Catheters

For some patients, effective pleurodesis will not be attainable. In others, the life expectancy could be too short to justify such invasive procedures or minor surgeries (such as VATS), but control of the effusion is still important. Therefore, other effective means for the palliation of pleural effusions are necessary. This need ultimately led to the development of the Denver Pleurx system (Denver Biomaterials Inc., Golden, CO), the only tunneled catheter approved by the U.S. Food and Drug Administration (FDA) for pleural effusion management.[96]

The details of the Denver catheter are described in detail elsewhere, but it consists primarily of a 15.5 Fr silicone catheter with side holes and a polyester cuff that induces fibrosis along the tunnel; fibrosis decreases the risks of dislodgement, infection, and pericatheter leakage.[96] There is a proximal hub that prevents air entry or fluid egress, and it is capped when not in use. The catheter is usually placed into a free-flowing effusion or large locule under local anesthesia at the bedside. After initial drainage of no more than 1500 mL, drainage usually is performed every other day by the patient, family member, or a visiting nurse. The tube may be removed if three drainages in a row are scant and imaging shows no reaccumulation of fluid, suggesting spontaneous pleurodesis. The pivotal, approval trial for the FDA in 1999 compared the Denver catheter with doxycycline sclerotherapy.[97] Quality-of-life benefits were the same for the two groups, and there was a greater improvement in dyspnea with the Denver catheter; hospitalization was shorter in the catheter group (1 day vs. 6.5 days), and spontaneous pleurodesis developed in 46% of catheter placements, but effusions did recur 13% of the time. Sclerotherapy may be given through the catheter as well. Catheter complications occur about 15% to 20% of the time and include poor drainage requiring replacement, external catheter migration, tumor tracking along the catheter route, and infection (pleural fluid and skin). A recent report demonstrates symptomatic benefit (but not pleurodesis) for 91% of patients with trapped lung or multiloculated effusions.[96] The Denver Pleurx system is likely to offer significant benefit to all patients with pleural effusion and might allow many stable patients to undergo drainage and sclerosis on an outpatient basis.

Approach to Management

This subject is not without controversy, particularly with respect to the method of pleurodesis. Surgeons are likely to favor a surgical approach, such as VATS or thoracoscopy with pleurodesis attempts using talc or pleural abrasion, whereas pulmonologists and oncologists are likely to favor chest tubes and talc. Newer approaches with smaller chest tubes allow for outpatient management of pleural effusions in a more stable patient population. Cost considerations are important as well; surgical approaches can add significant expense because of operating room time and the involvement of anesthesiologists, although some of these expenses can easily be recovered by quicker hospital discharges and elimination of complications. Certain physicians have strong feelings one way or the other based on their interpretations of the literature and their personal experiences; well-performed clinical trials are difficult to conduct in these situations and are unlikely to occur. For example, because of the low, but very real incidence of ARDS (and mortality) associated with the use of talc that has not been seen with the use of any other agents, some clinicians feel that intrapleural talc should never be used; others recommend talc as long as lower doses (2–5 g) are used, although ARDS can be associated even at these doses. [76] [78] [79] [82] Because many patients achieve pleurodesis through the presence of a chest tube alone, that sole measure might be sufficient for a good percentage of patients, and certainly a trial of this can be considered in stable patients. Although one general approach is shown in the algorithm, the overall scheme needs to be individualized for the patient, with one physician (usually the medical oncologist) as the coordinator to ensure optimal outcomes.


Malignant pericardial effusions are much less common than both ascites and pleural effusions. They are often an event that occurs with end-stage disease, and patients are often switched to a purely palliative mode on discovery of a malignant pericardial effusion, particularly if tamponade or near-tamponade physiology is present. Because primary intracardiac tumors are exceedingly rare, the most common cause of pericardial effusions is tumor metastatic to the heart or pericardium. About one-third of patients with end-stage lung or breast cancer have evidence of pericardial metastases and/or effusion present at autopsy, although only a small number of these are symptomatic. Figure 60-2 provides a treatment approach to malignant pericardial effusions.


Figure 60-2  Treatment approach to malignant pericardial effusions.



Etiology and Pathogenesis

The pericardial space normally contains a tiny amount of fluid that serves to reduce friction, maintain position of the heart in the pericardium and mediastinum, and provide a barrier against infection. Malignant cells obtain access to the space either by direct invasion from an adjacent tumor in the lung or mediastinum or by hematogenous or lymphatic spread. Retrograde progression of disease through the lymphatic channels draining the heart and pericardium results in a majority of effusions.[98] The potential space in the pericardial cavity is much less than in other potential areas of fluid accumulation (abdomen, lungs), and, depending on the compliance of the pericardium and the acuity of fluid collection, symptoms can develop rapidly. Furthermore, because even a small amount of fluid that develops rapidly can compress the heart and thus impede diastolic filling, malignant pericardial effusions are often difficult to manage and tend to carry a grim prognosis.

Clinical Presentation

The most common clinical manifestations of pericardial effusion are dyspnea, cough, chest pain, orthopnea, palpitations, tachypnea, tachycardia, and edema. Physical examination might show signs of a low-output cardiac state (coolness of the extremities, diaphoresis), jugular venous distention, distant heart sounds, narrowed pulse pressure, a pericardial friction rub, and a pulsus paradoxus. Pulsus paradoxus can be measured at the bedside using a sphygmomanometer that is deflated very slowly. During deflation, the first Kortokoff sounds are heard only during expiration and subsequently throughout the respiratory cycle. The difference in systolic pressures at which the Kortokoff sounds are heard between inspiration and expiration quantifies the pulsus paradoxus, which is normally no more than 10 mm Hg. Classically, Ewart's sign (dullness at the left infrascapular area due to bronchial compression by a large effusion) may be seen, but it is rarely observed in practice.[99] Electrocardiography might show low-voltage complexes across all leads (especially when compared with a prior study) and electrical alternans. Electrical alternans is considered a pathognomonic sign of pericardial tamponade; it is caused by a pendulum-like swinging of the heart in a fluid-filled pericardial cavity, but it is seen in less than 3% of cases.[100] Low voltage is defined as a total amplitude of the QRS complexes in each of the limb leads of 5 mm or less. Sinus tachycardia is often seen as well but has many other causes.

Diagnosis and Evaluation

A patient with a pericardial effusion, like a patient with pleural effusion, should be classified according to the likelihood of the effusion being malignant. The causes of pericardial effusions are shown inTable 60-3 and should be considered for all patients, because benign effusions clearly portend much better prognoses than malignant ones. Most effusions among patients with advanced cancer will, in fact, be due to that malignancy. An autopsy series of 3314 patients found that cardiac metastases occur in 10% of patients dying of cancer.[101] Many patients with malignant effusions have known advanced cancer, and therefore, there is little diagnostic dilemma. In those patients in whom advanced cancer is unlikely, a more thorough evaluation should be done.

Table 60-3   -- Major Differential Diagnosis of Pericardial Effusion by Etiology


Malignancy (usually metastatic)

Myocardial infarction associated (Dressler's syndrome)


Myxedema (rare cause of tamponade physiology)

Trauma (penetrating or nonpenetrating)


Acute idiopathic

Rheumatic fever

Collagen vascular disease (systemic lupus erythematosus, rheumatoid arthritis, scleroderma, Wegener's granulomatosis)

Postsurgical (cardiac and intrathoracic)

Drug induced (procainamide, hydralazine, phenytoin, doxorubicin, isoniazid)


Viral (coxsackievirus, echovirus, mumps, adenovirus, hepatitis, human immunodeficiency virus)

Bacterial (pneumococcus, streptococcus, staphylococcus)


Fungal (histoplasmosis, coccidiomycosis, candida, particularly in immunosuppressed patients)



Unfortunately, malignant pericardial effusion is often difficult to definitively diagnose. [100] [102] Lymphatic obstruction by tumor might cause blockage without direct involvement of malignant cells in the pericardium, thereby making the usual gold standard test, cytology, unhelpful. Other causes such as prior radiation also can cause a pericardial effusion and should always be considered in a patient with such a history to avoid mistakenly classifying a patient as having advanced cancer.

Electrocardiogram and physical examination can be performed with particular attention to the aforementioned findings. Chest radiograph can show a water bottle-shaped pericardium or an enlarged silhouette. Echocardiogram should be performed, emergently if there are clinical signs of tamponade, and to assess the hemodynamic impact of an effusion; this test can detect a volume of fluid as little as 15 mL. Fluid first appears as a lucent space between the pericardium and epicardium, visible only during systole and behind the left ventricle; once reaching 25 to 50 mL, it may be seen throughout the cardiac cycle. Once the effusion can separate the pericardium from the epicardium, the pericardial image becomes stationary; once it extends behind the atrium, atrial wall motion is markedly increased. When tamponade physiology is reached, the echo-Doppler shows compression of the right side of the heart, increased respiratory variation of mitral and tricuspid inflow velocity, and right atrial and ventricular collapse during diastole. Assessment of tamponade should be based on clinical and imaging evidence, not just on the volume of fluid present. CT scan is also sensitive for diagnosing an effusion. It can detect as little as 50 mL of pericardial fluid and, similar to an echocardiogram, can give an idea of intracardiac masses.[103] MRI also can provide direct imaging of the pericardium.[104] Both of these tests can give some clues as to the nature of the fluid (bloody, serous, chylous), but they rarely provide clinically useful information. If there is a small, questionable effusion present, MRI can differentiate between a small effusion and epicardial fat. Because of the frequency with which oncology patients receive CT scans and MRIs, these studies often identify an effusion, prompting a further evaluation. Cardiac catheterization of the left and right sides of the heart can be used to confirm tamponade physiology (demonstrated by equivalence of pressures across all chambers) and improvement of these pressures after pericardiocentesis. However, this is usually not necessary.


Pericardiocentesis is usually performed with a 16- to 22-gauge needle (often a spinal needle) attached to a syringe inserted at roughly a 45-degree angle below the xiphoid process cephalad toward the tip of the left scapula. The needle is generally attached to an electrocardiograph machine during the procedure. Advancement into the myocardium usually reveals an injury pattern on the electrocardiogram. Although this procedure is usually performed semi-electively, pericardiocentesis occasionally must be done in an emergent setting, where removal of as little as 50 mL of fluid can improve hemodynamic status.[105] Complications include ventricular perforation, arrhythmias, and pneumothoraxes, and range from 5% to 20%. Complications are less likely (about 2%) when echocardiography is used to delineate the size and location of the fluid with respect to normal cardiac structures.[106] Among patients who are clinically unstable, with tamponade physiology and hypotension, vigorous volume resuscitation should be performed to increase cardiac filling pressures and cardiac output, even if signs of heart failure, such as edema or lung rales, are seen.[107]

Pericardial Fluid Evaluation

The evaluation of a patient with effusion depends on several characteristics. For those patients with small effusions and advanced malignancy found incidentally on imaging, nothing needs to be done except a good physical examination and monitoring. For patients with a suspicion of cancer, or those patients who are symptomatic due to tamponade physiology, attempts at diagnosis and therapy should be made. Pericardial effusion is rarely the initial manifestation of malignancy. Two studies demonstrated that 7% of patients presenting with a pericardial effusion were ultimately given a new diagnosis of malignant disease.[108] Patients with no history of cancer who present with an effusion and absence of tamponade or significant clinical symptoms should be managed in a cautious, noninvasive manner. This approach usually includes treatment with anti-inflammatory therapy (such as nonsteroidal anti-inflammatory drugs) for potential idiopathic pericarditis for a period of 1 to 2 weeks.[108] Such management should be undertaken in cooperation with an experienced cardiologist. In terms of evaluating pericardial fluid, however, there are no specific guidelines for classifying an effusion as a transudate or exudate, or to further delineate the etiology of the effusion with criteria such as LDH, protein, glucose, pH, or cell count, as there are with pleural effusions. Interestingly, the best predictor for the behavior of a pericardial effusion of unknown etiology is from the clinical history of the patient. In a study of 322 patients from Sagrista-Sauleda and colleagues,[109] a large effusion with clinical signs of inflammation (two or more of the following: chest pain, fever, diffuse ST segment elevation, or friction rub) had a likelihood ratio of 20 for an idiopathic effusion, and those patients with tamponade without inflammation had a likelihood ratio of 2.9 for a malignant effusion.

It is worth repeating here that pericardial fluid, unlike pleural fluid, cannot be classified as a transudate or exudate. Nevertheless, aspirated fluid should be sent for cell count, cultures (if there is a possibility of infectious etiology), and cytology. The sensitivity of cytology for malignant pericardial disease varies widely, from 50% to 90%. [100] [102] Reasons given for false-negative cytology include limited cellularity in the specimen, shrouding by blood, or the absence of an expert cytopathologist. Interestingly, DNA diploidy obtained from flow cytometry has been found to correlate with benign cytology, and aneuploidy is associated with malignant cytology but is not sensitive enough to be definitive.[110] It is uncommon for pericardial fluid cytology to provide the first diagnosis of a malignancy. In one study of 47 pericardial fluid specimens, only 10 were positive for malignancy, and none of these represented the first diagnosis of a malignancy in a patient.[111] It is imperative, however, that the physician realize that a negative cytology does not exclude malignant pericardial disease. By the same token, patients with a history of cancer could have an effusion for other reasons, particularly radiation.[112] One should be cautious, then, in providing the diagnosis of a malignant effusion (and usually advanced, incurable malignancy) unless definitive cytologic evidence is seen. For patients with a negative cytologic specimen but suspected malignancy, pericardial biopsy may be performed. Prospective studies have demonstrated a 5% to 20% positive yield for diagnostic pericardial biopsy specimens.[108] This is probably due to the blind nature of the biopsies and to the practice of taking biopsy specimens from the parietal pericardium, whereas the principal site of malignant involvement is the visceral pericardium, site of the epicardial lymphatics. The positive yield is much higher (54%) for biopsies performed during therapeutic (subxiphoid pericardiostomies) procedures; although there might be some selection bias, therapeutic procedures allow the pericardium to be visualized more directly and thus permit more accurate biopsy site selection.[108]

Pericardioscopy is a relatively new technique available in specialized centers that allows endoscopic inspection, aimed biopsies, and drainage of both pericardial surfaces, and can even be used to obtain epicardial biopsy specimens.[113] An expansion of this technique, using a perDUCER catheter, can be used for endocardial biopsies and thus far has been used for diagnosing and managing patients with nonmalignant pericarditis. This technique, however, also allows for guidewire and catheter placement, by which intrapericardial therapy may be administered. Ultimately, this technique might be expanded for use in evaluating and managing malignant pericardial etiologies.[114]

Treatment and Management

Asymptomatic malignant effusions do not require therapy. Volume depletion should be avoided, because adequate right ventricular preload is essential to allow sufficient cardiac output. For symptomatic patients pericardiocentesis may be performed, either as initial therapy or to allow stabilization until a more definitive procedure can be performed. Unfortunately, as many as three quarters of patients have recurrence and require treatment.[115]


A limited number of patients might benefit from systemic chemotherapy if such options are available (i.e., if they are expected to be chemotherapy responsive), or if a delay in treatment increases the risk of significant systemic progression. Such chemotherapy-sensitive tumors include leukemias, lymphomas, germ cell tumors, and even breast cancer. Although there might be reluctance on the physician's part to treat a malignant pericardial effusion with chemotherapy because of the feeling that even moderate effusions could become life threatening, there is ample evidence in the literature to support the use of chemotherapy in many circumstances. There are reports of successful resolution of effusions for patients with chronic myelomonocytic leukemia with hydroxyurea or chemotherapy.[116] Vaitkus and associates[115] reviewed the experience of treatment of 46 patients, mostly with breast cancer or lymphoma, who had malignant pericardial effusions treated with chemotherapy (38 of 46 patients initially having undergone therapeutic pericardiocentesis), and reported that systemic chemotherapy prevented recurrence in 67% of these patients. A special entity, primary cardiac lymphoma, although it responds to traditional chemotherapeutic regimens such as CHOP (cyclophosphamide, doxorubicin, vincristine, and prednisone), is not responsive to radiotherapy and carries a poor prognosis, probably due to myocardial infiltration and tissue necrosis that might precede or result from chemotherapy.[117] Several researchers have reported pericardial effusions from leukemias with the presence of blast cells in the pericardial fluid, which were treated with emergent pericardiocentesis (because of tamponade) and immediate chemotherapy with good outcomes.[118] Other primary cardiac tumors, such as sarcoma and intrapericardial pheochromocytoma, tend not to be as responsive to chemotherapy, so control of the effusion should be the initial focus.[119] Vaitkus and coworkers[115] also reviewed the experience of 54 patients treated with radiotherapy and found that this procedure was successful two thirds of the time, mainly in cases of hematologic malignancies and breast cancer (93% and 71% success rates, respectively) although 45% of patients with other solid tumors also had successful control. It should be noted, however, that radiotherapy might induce scarring or fibrosis that makes further interventions more difficult, so this procedure should be saved for end-stage patients declining other interventions. Those patients with physiology demanding immediate intervention should have some sort of drainage performed, whereas those that are stable and potentially chemotherapy responsive should be considered for chemotherapy.

Percutaneous Tube Pericardiostomy

For patients not amenable to systemic therapy for whatever reasons, percutaneous drainage is performed. Because the rate of fluid reaccumulation is very high, prolonged catheter drainage is often used in addition to pericardiocentesis. This is now common practice, because it allows a more complete drainage than that attainable by pericardiocentesis alone. After the pericardiocentesis, a guide wire is left in place, and a pigtail catheter is placed via the Seldinger catheter-over-a-guide wire technique. The catheter is left in place for several days and is not removed until the drainage is less than 20 to 30 mL per 24-hour period, at which point the catheter is removed. There is a small risk of infection with a catheter remaining in place for several days.[120] The catheter should be subject to intermittent rather than continuous drainage to maintain catheter patency. With this technique, however, there is still a relatively high likelihood that the fluid will reaccumulate unless other steps are taken. Unless the patient has a very short life expectancy (which still might require a repeat drainage) or, in the physician's mind, the patient has a tumor from which one would expect a response from chemotherapy, other interventions will be required.

Surgical Approaches


The subxipoid pericardial window, first reported by Napoleon's surgeon, Larrey, in 1829,[121] is now the most common surgical procedure for pericardial effusions. The procedure may be performed under local anesthesia with intravenous sedation; a small incision is made from the xiphoid process caudally for approximately 5 cm, and the xiphoid is either bisected or resected. After dissection to the pericardium, adhesions and masses are identified, a 2- to 4-cm2 piece of pericardium is resected, and a drainage tube is placed. The drainage tube remains in place for several days, which could allow for sclerosis simply by acting as a pericardial irritant.[122] Allen and colleagues[123] retrospectively compared the experience of 117 patients who underwent either percutaneous catheter drainage with ultrasound or fluoroscopic guidance (n = 23), or subxiphoid pericardiostomy (n = 94). Complication rates were 17% in the percutaneous drainage group and 1% in the pericardiostomy group, with 4% mortality and a 30% recurrence rate in the percutaneous drainage group, compared with 0% mortality and 30% recurrence in the pericardiostomy group. In the percutaneous group, pigtail catheters remained in place for an average of 4.2 days, and mediastinal drainage was maintained for 5 days in the surgical group, so there was minimal difference in the number of days hospitalized. The advantages of the surgical group come with the caveat that the 23 patients who underwent percutaneous drainage did so because they were deemed to be too hemodynamically unstable to undergo a surgical procedure, and thus bias could account, at least in part, for the advantages of subxiphoid pericardiostomy. For all patients in this study, the median survival time of those patients having a malignant pericardial effusion was a paltry 2.2 months, with 13.8% of patients having a 1-year survival. No patients who underwent pericardiostomy developed constrictive pericarditis. For patients who are hemodynamically stable, the subxiphoid pericardiostomy can and should be considered in lieu of percutaneous catheter drainage.


Before the revival of the subxiphoid pericardial window, alternative surgical techniques had been used, including sternotomy and pericardiectomy or creation of a window via thoracotomy or VATS. [124] [125] Initially, the experience of Piehler and associates[124] with 145 patients suggested that the extent of pericardial resection affected the recurrence rate, thus favoring open surgical drainage rather than the easier subxiphoid pericardiostomy. Several recent studies, however, have demonstrated no difference between the open or VATS procedures and subxiphoid pericardiostomies. [126] [127] Vaitkus and coworkers[115] reported a decreased number of further pericardial complications when subxiphoid pericardiostomies instead of open pericardiectomies were performed. There is a much lower incidence of postoperative complications (including pneumonia, thrombosis, arrhythmia, respiratory failure) after subxiphoid pericardial drainage compared with transthoracic pericardial resection (10% vs. 50%, respectively). [126] [127] VATS offers a minimally invasive technique for treatment of effusions but still requires general anesthesia and the ability to tolerate single-lung ventilation, so it offers little advantage over subxiphoid pericardiostomy. It should therefore be saved for situations in which the subxiphoid approach fails, in which a VATS approach would assist with evaluation or treatment of simultaneous lung pathology, or in which a larger specimen of pericardium is required for clearer observation and evaluation of an undiagnosed effusion. [127] [128] A pericardial-peritoneal shunt with a Denver-type catheter and pump, similar to that used in pleural effusions, has been used for some patients with success, but its use should be limited to patients whose effusions are refractory to other management techniques.[128]


Percutaneous balloon pericardiostomy is an extension of a percutaneous pigtail catheter, except that a balloon-dilating catheter (20 mm diameter and 3 cm length) introduced over a guide wire creates a nonsurgical pericardial window. Ziskind and colleagues[129] provided the results from the first 50 patients treated with this technique. The procedure was considered successful in 46 of the 50 patients, with two patients requiring early operation for a bleeding vessel and persistent drainage and two patients requiring late drainage for recurrent tamponade. Minor complications included fever, thoracentesis, requirement for chest tube placement, and pneumothorax. The most significant complications were the development of pleural effusions requiring drainage. Subsequent studies have reported up to 100% success rates. [130] [131] Long-term outcome from these patients was still poor, with a mean survival time of 3.3 months. The procedure has been slow to gain acceptance because of the need for specialized training and equipment but could become a more popular option as interventional radiologists become adept in this technique.


For pericardial effusions, as for pleural effusions, sclerosing agents may be used, and in general, do work.[132] Maher and associates[133] reported their experiences with 93 patients who underwent pericardial fluid drainage, followed by sclerosis with either doxycycline or tetracycline. A median of three instillations was required and was able to control the effusion 88% of the time. Shepherd and colleagues[134] reported similar results (75% success rate). The most common complaints with this procedure are pain (intrapericardial lidocaine should be injected before the sclerosing agent) and arrhythmias. Bleomycin has been compared with doxycycline in a prospective, randomized study showing equal efficacy and less pain; therefore, it should probably be considered instead of doxycycline as the sclerosing agent of choice.[135]

The use of direct intrapericardial instillation of chemotherapy and biologic therapy has been examined in several settings with limited data. For pericardial as for pleural effusions, it is somewhat difficult to determine whether intrapericardial chemotherapy functions because of tumor cytotoxicity or pericardial irritation leading to sclerosis, because the physiology and mechanisms of sclerosis are poorly understood. Imamura and coworkers[136] looked at the use of OK-432 (a heat-treated lyophilate of Streptococcus pyogenes A3) instilled directly into the pericardium after fluid drainage. All 10 patients in this study achieved complete control of the pericardial effusion, 7 of the 10 with only one treatment, for an average of 10.8 months. Complications were primarily fever and chest pain. Hypotension was also seen and was thought to be a result of rapid fluid reaccumulation in several patients. Moriya and associates[137] explored the use of intrapericardial carboplatin in 10 patients with advanced non-small-cell lung cancer; 8 patients showed complete regression of the effusion with minimal toxicities and minimal systemic distribution of the carboplatin. Cisplatin is the best-studied therapy, with overall success rates of 67% to 87%.[138] One of the earliest studies on the use of cisplatin comes from Fiorentino and colleagues, who gave five patients intrapericardial cisplatin (50 mg over 5 minutes 5 days in a row, with courses repeated every 2–3 weeks if there was fluid recurrence).[138] There was a 60% complete response rate in the effusion. Side effects were limited to mild nausea with no hematologic or renal toxicities. Kawashima and colleagues[139] gave intrapericardial aclarubicin to five patients with pericardial tamponade secondary to malignant effusions; in four of the five, physicians were able to remove the pericardial catheter, and two of the five patients experienced a complete remission of the pericardial effusion. All patients demonstrated disappearance of the malignant cells from the pericardial space, with no cytopathologically demonstrable pericardial recurrence. Mitoxantrone and interferon have also been used, with similar results. [140] [141] With intrapericardial use, however, there is a significant risk of pain, a requirement for multiple instillations, and a possibility of constrictive pericarditis for patients with a prolonged survival. Because of these limitations the use of pericardial sclerosis is declining, and intrapericardial chemotherapy remains highly experimental.


Malignant pericardial effusion is a dreaded complication of malignancy, indicative of advanced disease in most cases. The oncologist should take care to ensure that advanced disease is in fact present, and that the effusion has no other etiology in a patient who has a history of cancer. Management should be focused on immediate resuscitation and control of life-threatening symptoms. Asymptomatic patients with cancer and an effusion can be managed conservatively and observed with either repeat careful physical examination or serial echocardiograms. If there is a strong likelihood of tumor response to chemotherapy, or if chemotherapy is absolutely required (e.g., in cases of lymphoma or life-threatening systemic disease), it should be started as soon as possible. If emergent pericardiocentesis is required, it should be performed with placement of a percutaneous pigtail catheter. For patients who are clinically more stable but unable to be treated adequately with systemic chemotherapy, one could consider either percutaneous drainage or subxiphoid pericardiectomy under local anesthesia, with probably similar results; the latter demonstrates a lower rate of recurrence and is the procedure of choice. Such a management approach has been validated in several studies. [123] [142] Use of other techniques, such as VATS, percutaneous balloon pericardiotomy, sclerosis, or intrapericardial chemotherapy offers new potential options for these patients should primary therapy fail.


Malignant ascites is a frequent occurrence in the patient with malignancy and is usually an indication of peritoneal carcinomatosis. Like malignant pericardial and pleural effusions, it is often a harbinger of advanced disease. It accounts for 10% of cases of ascites, and 1-year survival is less than 10%.[143]

Etiology and Pathogenesis

The peritoneal membrane contains lymphatics, which serve to collect fluid, proteins, cells, and other substances and return them to the systemic collection. There is a specialized anatomic feature, the stomata, which are relatively open connections that connect the abdominal cavity to submesothelial lymphatics and that are the sites of most lymphatic drainage of the abdomen.[144] Fluid accumulations, as in the pleural and pericardial cavities, can occur if there is overproduction or decreased drainage of fluid. Studies in adults have demonstrated that increased fluid production is a major cause of ascites among patients with peritoneal carcinomatosis.[145] Animal models in which mice were given intraperitoneal injections of tumor cells have demonstrated that obstruction to lymphatic flow can cause ascites as well, and this has been confirmed using lymphoscintigraphy in humans. [146] [147] VEGF, an angiogenesis factor expressed by most tumors (including those that spread to the peritoneum), has been demonstrated in vivo to cause increased fluid production and to influence the development of ascites. [148] [149] [150] VEGF is probably the single biggest mediator of ascites production. Interleukin-2, tumor necrosis factor, interferon-a, and matrix metalloproteinases have also been implicated as factors contributing to the formation of ascites. [151] [152] [153] Intraperitoneal protein accumulation decreases the difference between plasma and peritoneal oncotic pressures and subsequently decreases filtration into the lymphatics.[154] Free fluid flow into the abdomen therefore increases. Fluid protein concentration can reach oncotic pressures similar to that reached in peritoneal dialysis, which is very effective in causing ultrafiltration.[155] The most common tumors that cause ascites are those of gastrointestinal and pelvic organs; the single most common is ovarian cancer.

Diagnosis and Evaluation

Many patients with malignant ascites have small amounts of fluid first noticed on imaging studies, such as those done for staging or evaluation after therapy or for another complaint. Symptoms resulting from large amounts of fluid usually include abdominal distention (such that trousers no longer fit properly), early satiety, nausea, vomiting, an increase in weight, edema, and shortness of breath. Physical examination in the evaluation of ascites is notoriously difficult, especially among obese patients with lesser amounts of fluid. The physical examination has been shown to be from 50% to 94% sensitive and from 29% to 82% specific, with ultrasound being the gold standard.[156] Physical examination can demonstrate a fluid wave, shifting, or flank dullness. The absence of flank dullness is the most accurate predictor against ascites, but one requires 1500 mL of fluid to be present to notice flank dullness. If there is suspicion of ascites in any patient, ultrasound is the easiest, fastest, and cheapest method of ascertaining a diagnosis of ascites; one series demonstrated that as little as 100 mL of fluid can be demonstrated on an ultrasonogram.[157] Radiologic studies such as CT or MRI can also give clear-cut evidence of ascites, but these are usually not necessary as part of the first-line evaluation. Plain-film radiograms can show a ground-glass appearance to the abdomen, loss of detail, haziness, or floating bowel on supine films but is nonspecific. Figure 60-3 gives a treatment approach to malignant ascites.


Figure 60-3  Treatment approach to malignant ascites.



For many patients with known malignancy, the presence of ascites will not alter management, because advanced disease is already present. For some patients, particularly those without a known malignancy, the detection of ascites requires an appropriate evaluation to determine the etiology. For these patients, the differential diagnosis includes CHF, alcoholic cirrhosis, cirrhosis due to hepatitis B or C, previous abdominal surgeries, nonalcoholic steatohepatitis, and other causes of increased portal venous pressure. It should be remembered that patients with a history of cirrhosis have a predilection for hepatocellular carcinoma.

In most cases of ascites in which malignancy is a possibility, cross-sectional imaging, such as with a CT scan, should be performed to seek a primary tumor site, hepatic metastases, or other evidence of peritoneal seeding. Women in particular should have either an ultrasonogram or CT scan to evaluate the ovaries. After this, a diagnostic abdominal paracentesis is indicated. Ultrasound may be used during the procedure to localize ascites and minimize the risk of injury to abdominal organs (particularly the liver and intestine).[157] After local anesthesia, an 18-gauge needle attached to a 25- to 50-mL syringe is inserted into a site indicated by ultrasound or into the areas percussed as dull in the flanks, if sufficient fluid is present. If cytology is required, several liters of fluid can be obtained using vacuum bottles. Even among patients with coagulopathies (such as cirrhosis), the risk of developing a hematoma is only 1% and that of causing peritonitis or hemoperitoneum is 0.1%; therefore, one can usually proceed without the use of clotting factors or platelets.[158] Patients on therapeutic anticoagulation should have their anticoagulation withheld for several days before paracentesis.

Visualization of the fluid can provide some useful clues; clear fluid is usually associated with cirrhosis. Infected fluid is cloudy. Milky fluid can indicate chylous ascites and should be sent for triglyceride evaluation. Such fluid often has triglyceride levels greater than 200 mg/dL and often as high as 1000 mg/dL. Some studies have demonstrated that the most common cause for chylous ascites is malignancy, although others have found cirrhosis as the primary cause. [159] [160] Either way, if chylous fluid is found it should heighten the physician's attention to the potential for malignancy. Heterogeneously bloody fluid is associated with a traumatic tap, whereas homogeneously bloody fluid indicates prior bleeding (i.e., a nontraumatic tap) and could indicate malignancy. Ascites is bloody in up to one half of patients with hepatocellular carcinoma and in 22% of all patients with malignant ascites.[161] Spontaneous bacterial peritonitis (SBP) does not tend to develop to the extent that it does among patients with cirrhotic ascites, unless there has been previous surgery or a previous paracentesis. All patients should be evaluated for signs of SBP (including fever and abdominal pain), and a cell count on the ascitic fluid should be done; a polymorphonuclear cell count greater than 250 per cubic millimeter should prompt the use of empiric antibiotics in appropriate situations.[158]

The first and most important step in the evaluation of the patient with ascites of unknown etiology is to differentiate those causes arising from portal hypertension (usually cirrhosis) from other causes (including malignancy). The best test for this determination is the serum-to-ascites albumin gradient, which is the difference between serum albumin and ascitic fluid albumin. Low-protein ascites is usually from portal venous hypertension. A gradient equal to or greater than 1.1 g/dL indicates portal hypertension with 97% accuracy, whereas a lower gradient (high-protein ascites) indicates a lack of portal hypertension and possibly the presence of a malignancy.[160] The classical division of ascitic fluids into transudates and exudates using LDH, protein, and their serum-to-ascites ratios, akin to pleural effusions, has been shown to be not as useful. Total protein is not particularly useful for diagnosis of malignant ascites, although a value of less than 1 g/dL is useful in predicting a patient's increased risk for SBP.[162] This range of values is usually seen in cirrhotic patients. A glucose value below 50 mg/dL has been associated with malignancy but also can be indicative of infection.[27] An ascites-to-serum ratio of LDH greater than 1 indicates that the enzyme is actively being produced in the ascitic fluid and suggests malignancy, but not as specifically as with a pleural effusion. Triglyceride levels should be obtained in milky ascites but might or might not be particularly helpful. Other chemistries and cultures should be obtained if diagnoses other than malignancy are entertained. [143] [158] Siddiqui and associates[163] have demonstrated that fibronectin is up to 100% sensitive and specific as a marker of malignant ascites in a small study of only 12 patients with malignant ascites; others feel that this test is not helpful, and it is not used routinely.[164]

The detection of tumor cells by cytology remains the gold standard for the detection of malignancy. For patients with peritoneal carcinomatosis due to cellular exfoliation into the ascitic fluid, malignant cells can be detected nearly 100% of the time.[161] On the other hand, patients with liver metastases and portal venous obstruction, chylous ascites from lymphomas and lymphatic obstruction, or hepatocellar carcinoma, might not have a positive ascitic fluid; the overall sensitivity for cytology is therefore on the order of 40% to 75% when 500 mL of fluid is obtained. [163] [165] To increase the detection of tumor cells and differentiate mesothelial cells from malignant cells, immunohistochemistry for cytokeratin, vimentin, leukocyte common antigen, S100, CEA, HMB45, and other markers may be performed on the cytologic specimen. A small study by Loewenstein and coworkers[166] has shown that an elevated CEA level in ascitic fluid is a good marker for malignancy, although this has not been well validated and requires further evaluation. Elevated serum or ascites CA-125 and CEA levels certainly should prompt a careful evaluation for malignancies, particularly for ovarian cancer in women with highly elevated CA-125, although neither marker is specific for any particular malignancy (or even malignancies in general) and might reflect only intra-abdominal pathology rather than malignant disease.[154] No known studies have correlated levels of CA-125 and CEA with the likelihood of a malignancy, but one would expect higher levels of tumor markers to have a greater association with malignancy.

Women with only ascites and no evidence of primary malignancy on cross-sectional imaging present an interesting situation. These patients should undergo laparoscopic evaluation, at which time biopsy samples are taken from both ovaries, ascitic fluid is sampled, and random biopsies are performed throughout the abdomen. If ovarian carcinoma or adenocarcinoma consistent with ovarian carcinoma is seen, the procedure is converted to a laparotomy; it has been shown that aggressive surgical debulking to minimal residual disease improves response, survival time, and response to aggressive chemotherapy (usually taxane- and platinum-based) among patients with ovarian carcinoma. Such an approach can lead to a long disease-free interval and significant long-term survival. Even small, low-grade ovarian tumors, with the primary tumor not seen on imaging, can lead to diffuse peritoneal contamination, and hence, to the development of ascites. [167] [168] Some patients might have papillary serous carcinoma of the peritoneum, which arises from the peritoneal surface but shares a common histogenesis with ovarian tissue and also might be difficult to detect on cross-sectional imaging. It is important to remember that patients with ovarian cancer with ascites are classified as only stage III, and hence this rather large subgroup should be treated aggressively whenever possible. Papillary serous carcinoma of the peritoneum should be treated in a fashion similar to ovarian carcinoma.[169] Laparoscopy should be performed diligently. In a study from van Dam and colleagues[170] of 104 women with advanced ovarian cancer who underwent laparoscopy for diagnosis, 58% of patients had tumor implanted at the trocar site when only the skin was closed at the conclusion of the procedure, but in those in whom all abdominal wall layers were closed, only 2% had such implantation. For patients with ascites, there is a risk of prolonged fluid drainage at trocar placement sites. In one study, 2 of 92 patients developed peritonitis and died after undergoing laparoscopic procedures in these circumstances.[171] For a more detailed discussion of the diagnosis and management of ovarian carcinoma, the reader is referred to Chapter 93 .


As with pleural and pericardial effusions, it is important to distinguish the appropriate situations in which to treat malignant ascites. Small amounts of ascites can be tolerated very well or not even noticed. On the other hand, when symptoms such as dyspnea, abdominal pain, fatigue, anorexia, or early satiety arise, treatment can be required. The most worrisome sign is dyspnea, which can arise either from an inordinate amount of weight being carried by the patient or by compression of the pleural cavity leading to a decrease in total lung capacity. Clearly, the overall clinical status of the patient should be taken into account when making these treatment decisions. In light of the fact that ascites does not usually pose a life-threatening situation, and that the disease might respond well to systemic therapy in instances such as ovarian cancer, systemic chemotherapy can be considered front-line therapy. For other situations, such as pancreatic cancer, because responses to therapy are much less frequent, one would consider an early direct approach to control of the fluid. Certainly, in either case, because most techniques to manage fluid (e.g., diuresis, paracentesis) are relatively easy and performed on an outpatient basis, one could combine systemic chemotherapy with these other options.

Fluid Balance Management

Intravascular fluid that leaks into nonvascular compartments, also known as third spacing, results in decreased renal blood flow and consequently to sodium and water retention by the kidney, which ultimately leads to the propagation of further third-spacing of fluids. The use of fluid management techniques to manage malignant ascites is controversial. Because tumors that arise from peritoneal seeding are not caused by pressure gradient shifts, some feel that management using diuretics and salt management techniques that affect renal handling of excess fluid and sodium are useless, whereas others believe they can be effective. [172] [173] One can attempt dietary restriction of sodium to less than 2 g/day and also restrict free water intake to help decrease the amount of ascites. Unfortunately, to the patient with advanced malignancy, this strategy can prove overly burdensome, particularly when the patient is undergoing other treatments such as chemotherapy. At a minimum, however, the physician should restrict intravenous hydration as much as possible when chemotherapy is being administered.

The next step in this mode of management is the use of diuretics, which offers a noninvasive mechanism to help maintain fluid balance. The advantage of this approach is that it can be done easily on an outpatient basis with oral medications, and it is even possible for patients to perform minor self-adjustments in medication after instruction, which also restores to patients a degree of control over their own health. A distal tubule diuretic such as spironolactone is usually used first. Such agents act on the distal nephron, thereby minimizing the chances for compensation by more distal renal elements. More important, it works at the site of the renin-angtiotensin-aldosterone axis, which is upregulated due to decreased renal perfusion when significant third-spacing occurs, as in ascites. Doses start at approximately 25 to 50 mg/day and can go as high as 200 mg/day. Painful gynecomastia can result, and if this proves bothersome, amiloride is an alternative. Spironolactone alone often is insufficient and not immediate in its effects. In such cases, furosemide (usually starting at 20 mg/day and titrated to much higher doses) has been used for either a short period or, commonly, in combination. Razis and associates,[174] along with others, however, have pointed out that loop diuretics such as furosemide are minimally effective, if at all, in the management of ascites. The maximal ascitic reabsorption rate is 930 mL/day, and patients should therefore be instructed to weigh themselves daily and allow no more than a 0.5 to 1 kg/day loss in weight; in the presence of peripheral edema, slightly more weight can be lost per day.[144] Overdiuresis can result in hypotension, volume depletion, azotemia, and electrolyte abnormalities (particularly hyperkalemia with spironolactone, hypokalemia with furosemide), and thus patients should be monitored closely. Patients should be instructed to monitor their weights between physician visits and even to modify the doses of diuretics used when certain weight goals are met.

Unfortunately, although diuretics initially are used by many physicians for the control of ascites and certainly are easy to administer, they are not very effective.[175] Malignant ascites is usually not very responsive to diuresis because the pathogenesis is usually not related to increased portal pressure, as it is in cirrhosis, but rather is related to increased fluid production from the presence of tumor cells in the peritoneum.[176] In the rarer cases in which massive hepatic metastases are present and portal hypertension is the etiology of ascites, patients are much more likely to respond to diuretics.[177] Pockros and coworkers[176] confirmed this, finding that patients with a large amount of hepatic metastases (and hence elevated portal venous pressure) had a serum-to-ascites-albumin gradient similar to cirrhotics and were more likely to respond to diuretics, whereas those patients with peritoneal carcinomatosis were less likely to respond to diuretic management of ascites. Although there are no direct comparative studies, it is because of these clinical experiences that diuresis is used much less frequently than paracentesis as a means of controlling ascites.[175]


Paracentesis is the most frequently used and most effective management approach to malignant ascites, but its effects are only temporary.[175] Large-volume paracentesis can improve shortness of breath and early satiety quickly. Because one removes the fluid and not the cause of the fluid, however, there is rapid redistribution to the peritoneal cavity. Because of this, large-volume paracentesis can result in intravascular volume depletion, hypotension, azotemia, and other consequences of dehydration.[144] Colloidal volume expansion has been tried as a means to prevent these sequelae. In a randomized trial in cirrhotics by Gines and colleagues,[178] 105 patients receiving large-volume paracentesis were randomly assigned to receive albumin or not. Patients not receiving albumin were more likely to show signs of hemodynamic deterioration, worsening renal function, and hyponatremia (20.8% vs. 3.8%). Most of these patients, however, simply had laboratory abnormalities, because no advantages in clinical morbidity or survival ever have been demonstrated in this or any other study.[158] Further, a 5-L paracentesis would require approximately 50 g of albumin at a cost of up to several thousand dollars and would change a quick procedure into either a day-long one or one requiring an overnight hospital stay, which is not easily justified for patients with advanced malignancies in the absence of proven benefit.[7] Because of this, the use of any colloidal expanders is difficult to justify.

Complications of repeated paracentesis include bleeding, pain, the induction of peritonitis, and bowel perforation. The presence of loculated ascites can make adequate removal of fluid to provide a symptomatic benefit nearly impossible, and it can increase the risk of the procedure. Ultrasound guidance should be used to decrease the risk of complications.

Peritoneovenous Shunting

Peritoneovenous shunting, introduced by LeVeen for alcoholic liver disease, is effective in malignant ascites.[179] Because patients with intractable and debilitating ascites have a median survival of only 6 to 33 weeks and could probably be managed more easily with repeated paracentesis, patients selected for shunting should have an expected survival of several months requiring frequent paracentesis, as demonstrated by repeated recurrence of the ascites after drainage.[180] The device consists of a long, perforated tubing inserted in the peritoneal cavity, a tubing that inserts into the superior vena cava, and a one-way valve that connects the two. There are two types of shunts, the LeVeen shunt and the Denver shunt; the latter has a one-way pump that can be used by the patient or physician to clear debris in the shunt or valve. On inspiration, intrathoracic pressure is lowered and the one-way valve opens, allowing a baseline pressure difference between the thoracic cavity and abdominal cavity to increase by 3 to 5 cm H2O from its baseline difference of 5 to 15 cm H2O, causing fluid to flow across the pressure gradient and drain into the superior vena cava. Placement of the shunt can be done under local anesthesia, with two to three separate incisions made. Parsons and associates demonstrated no survival or quality-of-life advantage when peritoneovenous shunting was compared with repeated paracentesis.[177]Peritoneovenous shunting is effective in controlling ascites between 62% and 88% of the time. [181] [182] Early studies demonstrated a large number of pump failures due to occlusion (up to nearly two thirds), but more recent studies have demonstrated lesser rates of occlusion. [179] [181] [183] The manual pump present in the Denver device for clearing blockages has not demonstrated an advantage in the maintenance of shunt patency.[181] Flushing the pump and administration of thrombolytic agents might be able to restore patency and avoid pump removal and replacement in instances of pump failure.

Patients with ascites that is cytologically negative for tumor cells have a much longer shunt life than those patients with cytologic evidence of malignancy; this is probably due to sludging of the pump system with more viscous tumor cells or debris in the latter case. Immediately after placement of the pump, CHF can result from a rapid increase in intravascular volume from infusion of a large amount of ascitic fluid; this risk can be minimized by performing a large-volume paracentesis just before the procedure. Coagulopathy—specifically, disseminated intravascular coagulation (DIC)—is common when the pumps are used in cirrhotic patients, but not in those with malignant ascites (rate of 4%).[184] DIC probably arises as a result of fibrinolytic activity of ascitic fluid, and fibrinolytic activity is decreased in patients with malignant ascites.[185] Those patients with malignant ascites and good preshunt hepatic function seem to be at an even particularly low or negligible risk for DIC. [186] [187] Widespread dissemination of tumor cells from the peritoneum to the lungs has been demonstrated in several studies. [179] [183] [187] The clinical significance of additional tumor emboli among patients with advanced refractory malignancy, although it represents a major potential complication of shunt placement, is unclear. Infection also remains a concern, but the incidence of shunt-induced peritonitis is much lower among patients with malignant disease than among those with cirrhosis, probably because of the higher levels of protein and immunoglobulin in the ascites.[162]

With significant potential complications and alternatives available, Souter and coworkers[183] suggest the following criteria for patients who should be considered for shunt placement:



The goal of care is palliative.



Expected survival is longer than 3 months.



The rate of fluid reaccumulation is rapid after large-volume paracentesis.



There is no loculation.



Accumulated fluid is not bloody or viscous, which could lead to early shunt dysfunction.

Patients with peritonitis or those not able to handle large, rapid fluid shifts (patients with significant cardiac or renal dysfunction) would also not be candidates for shunt placement.

Drainage Catheters

External drainage catheters offer a different method for palliation of ascites, namely a route available for repeated drainage that does not require repeated needle insertion. Patients are therefore offered the ability to perform repeated paracentesis without increased morbidity and discomfort, possibly at home by themselves or with minimal assistance. Several different methods are available; most use vascular ports inserted into the peritoneum, tunneled catheters, or other similar devices to allow drainage after catheter access. [188] [189] [190] The best-described uses of this method are from Lomas and colleagues,[191] in which a patient with malignant ascites had 1 L of fluid removed per day for 3 months with a Tenckhoff catheter, and that from Belfort and associates,[192] in which 17 patients were implanted with a 20 Fr silastic tube with a Dacron cuff at the peritoneal surface, all of whom did well with repeated removal of ascites.[193] A disadvantage of repeated drainage of ascites via any means is that, among patients who do survive and undergo a number of procedures, there is a large amount of protein loss, particularly when compared with peritoneovenous shunting, in which the fluid is rerouted to the vascular system and protein might be retained. The most impressive protein loss seen comes from the Lomas and coworkers study,[191] in which the average albumin level fell from 2.8 g/dL to 1.8 g/dL. Other series, however, have demonstrated much smaller decreases in albumin level, especially when high-protein diets were used. If one considers that these patients have advanced, refractory malignancy, protein and albumin loss due to repeated drainage is probably not all that significant but should still be monitored. [189] [192] There is a risk of infection with superficial catheters; in the aforementioned study by Belfort and colleagues,[192] in which patients were given cuffed catheters, 47% of patients (8 of 17) developed positive ascitic fluid surveillance cultures, and 12% (2 of 17) required removal due to “significant infection.” In a study using a tunneled catheter (the Denver Pleurx catheter), 0 of 10 patients developed infections requiring catheter removal.[189] Although no comparison studies have been done with peritoneal catheters, randomized studies involving other catheter systems clearly have demonstrated decreased infection rates when catheters had their cuffs placed subcutaneously in a tunnel rather than at the surface.[192] Percutaneous catheters therefore remain an option for some patients, although they are not used frequently.


Sugarbaker[194] has been an advocate of peritonectomy, in which various parts of the peritoneum, omentum, and some intra-abdominal organs are removed as a method of tumor cytoreduction (akin to debulking of ovarian cancer, which has now become a standard of care) in preparation for intraperitoneal chemotherapy. This approach has been used for patients with peritoneal carcinomatosis with the chemotherapy tailored toward the nature of the malignancy. Studies have shown modest success with this procedure in increasing survival time and in the prevention of the development of malignant ascites, but its use in the treatment of ascites has not been well evaluated. [195] [196] It is likely to be of minimal use, because patients with advanced, malignant ascites often have chemotherapy-refractory disease and have too much systemic disease to benefit from cytoreduction; removal of ascites alone as palliation can usually be accomplished through much easier means. The role of intraperitoneal chemotherapy in the prevention of the development of ascites is intriguing, however, and is discussed next.

Intraperitoneal Therapy

Direct intraperitoneal therapy has been tried for quite some time in an attempt to deliver higher doses of chemotherapy locally with minimal systemic absorption and distribution.[197] Intraperitoneal therapy has been used to treat both ascites and intra-abdominal malignancies, such as ovarian cancer. Patients with disease responsive to chemotherapy are most likely to benefit from this procedure. In this manner, the intraperitoneal administration of cisplatin has been well studied, because ovarian cancer is one of the most common intra-abdominal tumors with a predilection for causing ascites and is very responsive to systemic cisplatin chemotherapy. Simultaneous administration of intravenous thiosulfate might decrease systemic absorption of cisplatin. Several investigators have found that intraperitoneal administration of cisplatin is indeed effective for control of intra-abdominal ovarian cancer, with perhaps better efficacy than that given intravenously.[198] The complexity of these procedures, their limited availability, and the fact that many or most patients who present with intractable ascites have already undergone numerous treatments and are probably resistant to chemotherapy, limits the role for intraperitoneal therapy in the management of ascites except in some special situations such as ovarian cancer and, of course, in clinical trials. It is important for the role of intraperitoneal chemotherapy immediately after initial surgical exploration for malignancy to be examined further to determine whether it produces a survival advantage or a reduction in the rate of development of malignant ascites.


The level of aggressiveness that one undertakes in the management of ascites is highly dependent on the overall clinical status of the patient. For those patients with chemotherapy-sensitive tumors who are not heavily pretreated, chemotherapy should be an early step in management, particularly in cases of ovarian and hematologic malignancies. Although diuretic therapy probably does not have much of an effect, one can consider an early trial, in that it is relatively easy to carry out and can be particularly helpful for patients with increased portal pressures due to massive hepatic metastases. For those patients with a limited life expectancy but a large amount of ascitic fluid causing symptoms, paracentesis on an as-needed basis should be used, with the caveat that recurrence of the effusion is virtually assured and repeat paracentesis required unless other steps are taken. Intraperitoneal therapy of many types can be considered where available, particularly for ascites due to ovarian cancer or in the setting of a clinical trial. Patients who are likely to live a long time without chemotherapy-responsive tumors should be considered for a shunt or catheter drainage device, or for any available clinical trial.


  1. Noppen M: Normal volume and cellular contents of pleural fluid.  Curr Opin Pulm Med2001; 7:180-182.
  2. Noppen M, De Waele MR, Li R, et al: Volume and cellular content of normal pleural fluid in humans examined by pleural lavage.  Am J Respir Crit Care Med2000; 162:1023-1026.
  3. Miserocchi G: Physiology and pathophysiology of pleural fluid turnover.  Eur Respir J1997; 10:219-225.
  4. Sahn SA: Malignant pleural effusions.  Clin Chest Med1985; 6:113-125.
  5. Grove CS, Lee YC: Vascular endothelial growth factor: the key mediator in pleural effusion formation.  Curr Opin Pulm Med2002; 8:294-301.
  6. Cheng D, Lee YC, Rogers JT, et al: Vascular endothelial growth factor level correlates with transforming growth factor-beta isoform levels in pleural effusions.  Chest2000; 118:1747-1753.
  7. Ishimoto O, Saijo Y, Narumi K, et al: High level of vascular endothelial growth factor in hemor-rhagic pleural effusion of cancer.  Oncology2002; 63:70-75.
  8. Leuallen EC, Carr DT: Pleural effusion: statistical study of 436 patients.  N Engl J Med1955; 232:79-83.
  9. Tinney WS, Olsen AM: The significance of fluid in the pleural space: a study of 274 cases.  J Thorac Surg1946; 14:248-252.
  10. Salyer WR, Eggleston JC, Erozan YS: Efficacy of pleural needle biopsy and pleural fluid cytopathology in the diagnosis of malignant neoplasm involving the pleura.  Chest1975; 67:536-539.
  11. Hirsch A, Ruffie P, Nebut M, et al: Pleural effusion: laboratory tests in 300 cases.  Thorax1979; 34:106-112.
  12. Chernow B, Sahn SA: Carcinomatous involvement of the pleura: an analysis of 96 patients.  Am J Med1977; 63:695-702.
  13. Johnston WW: The malignant pleural effusion. A review of cytopathologic diagnoses of 584 specimens from 472 consecutive patients.  Cancer1985; 56:905-909.
  14. Anderson CB, Philpott GW, Ferguson TB: The treatment of malignant pleural effusions.  Cancer1974; 33:916-922.
  15. Colins JD, Burwell D, Furmanski S, et al: Minimal detectable pleural effusions. A roentgen pathology model.  Radiology1972; 105:51-53.
  16. Sahn SA: State of the art. The pleura.  Am Rev Respir Dis1988; 138:184-234.
  17. Stark P: The pleura.   In: Taveras A, Ferucci C, ed. Radiology: Diagnosis, Imaging, Intervention,  Philadelphia: Lippincott; 2003:454-460.
  18. Mergo PJ, Helmberger T, Didovic J, et al: New formula for quantification of pleural effusions from computed tomography.  J Thorac Imaging1999; 14:122-125.
  19. Davis SD, Henschke CI, Yankelevitz DF, et al: MR imaging of pleural effusions.  J Comput Assist Tomogr1990; 14:192-198.
  20. Koh DM, Burke S, Davies N, Padley SP: Transthoracic US of the chest: clinical uses and applications.  Radiographics2002; 22:e1.
  21. Lipscomb DJ, Flower CD, Hadfield JW: Ultrasound of the pleura: an assessment of its clinical value.  Clin Radiol1981; 32:289-290.
  22. Grogan DR, Irwin RS, Channick R, et al: Complications associated with thoracentesis. A prospective, randomized study comparing three different methods.  Arch Intern Med1990; 150:873-877.
  23. Tarn AC, Lapworth R: Biochemical analysis of pleural fluid: what should we measure?.  Ann Clin Biochem2001; 38:311-322.
  24. McVay PA, Toy PT: Lack of increased bleeding after paracentesis and thoracentesis in patients with mild coagulation abnormalities.  Transfusion1991; 31:164-171.
  25. Collins TR, Sahn SA: Thoracocentesis. Clinical value, complications, technical problems, and patient experience.  Chest1987; 91:817-822.
  26. Aguilar-Torres FG, Schlueter DP, Perlman L, Maskawa T: Subcutaneous implantation of an adenocarcinoma following thoracentesis.  Wis Med J1977; 76:S19-S21.
  27. Capizzi SA, Prakash UB: Chest roentgenography after outpatient thoracentesis.  Mayo Clin Proc1998; 73:948-950.
  28. Light RW, Macgregor MI, Luchsinger PC, Ball Jr WC: Pleural effusions: the diagnostic separation of transudates and exudates.  Ann Intern Med1972; 77:507-513.
  29. Costa M, Quiroga T, Cruz E: Measurement of pleural fluid cholesterol and lactate dehydrogenase. A simple and accurate set of indicators for separa-ting exudates from transudates.  Chest1995; 108:1260-1263.
  30. Keller RR: Once more: Light's criteria revisited.  Respiration2000; 67:11-12.
  31. Valdes L, Pose A, Suarez J, et al: Cholesterol: a useful parameter for distinguishing between pleural exudates and transudates.  Chest1991; 99:1097-1102.
  32. Heffner JE, Brown LK, Barbieri CA: Diagnostic value of tests that discriminate between exudative and transudative pleural effusions.  Chest1997; 111:970-980.
  33. Goldstein LS, McCarthy K, Mehta AC, Arroliga AC: Is direct collection of pleural fluid into a heparinized syringe important for determination of pleural pH? A brief report.  Chest1997; 112:707-708.
  34. Lesho EP, Roth BJ: Is pH paper an acceptable, low-cost alternative to the blood gas analyzer for determining pleural fluid pH?.  Chest1997; 112:1291-1292.
  35. Good Jr JT, Taryle DA, Maulitz RM, et al: The diagnostic value of pleural fluid pH.  Chest1980; 78:55-59.
  36. Sahn SA, Good Jr JT: Pleural fluid pH in malignant effusions. Diagnostic, prognostic, and therapeutic implications.  Ann Intern Med1988; 108:345-349.
  37. Yam LT: Diagnostic significance of lymphocytes in pleural effusions.  Ann Intern Med1967; 66:972-982.
  38. Martinez-Garcia MA, Cases-Viedma E, Cordero-Rodriguez PJ, et al: Diagnostic utility of eosinophils in the pleural fluid.  Eur Respir J2000; 15:166-169.
  39. Light RW, Erozan YS, Ball Jr WC: Cells in pleural fluid. Their value in differential diagnosis.  Arch Intern Med1973; 132:854-860.
  40. Starr RL, Sherman ME: The value of multiple preparations in the diagnosis of malignant pleural effusions. A cost-benefit analysis.  Acta Cytol1991; 35:533-537.
  41. Hsu C: Cytologic detection of malignancy in pleural effusion: a review of 5,255 samples from 3,811 patients.  Diagn Cytopathol1987; 3:8-12.
  42. Banales JL, Pineda PR, Fitzgerald JM, et al: Adenosine deaminase in the diagnosis of tuberculous pleural effusions. A report of 218 patients and review of the literature.  Chest1991; 99:355-357.
  43. Sassoon CS, Light RW: Chylothorax and pseudochylothorax.  Clin Chest Med1985; 6:163-171.
  44. van de Molengraft FJ, Vooijs GP: The interval between the diagnosis of malignancy and the development of effusions, with reference to the role of cytologic diagnosis.  Acta Cytol1988; 32:183-187.
  45. American Thoracic Society : Management of malignant pleural effusions.  Am J Respir Crit Care Med2000; 162:1987-2001.
  46. Hengge UR, Ruzicka T, Tyring SK, et al: Update on Kaposi's sarcoma and other HHV8 associated diseases. Part 2: Pathogenesis, Castleman's disease, and pleural effusion lymphoma.  Lancet Infect Dis2002; 2:344-352.
  47. Beauchamp HD, Kundra NK, Aranson R, et al: The role of closed pleural needle biopsy in the diagnosis of malignant mesothelioma of the pleura.  Chest1992; 102:1110-1112.
  48. Prakash UB, Reiman HM: Comparison of needle biopsy with cytologic analysis for the evaluation of pleural effusion: analysis of 414 cases.  Mayo Clin Proc1985; 60:158-164.
  49. Emad A, Rezaian GR: Closed percutaneous pleural brushing: a new method for diagnosis of malignant pleural effusions.  Respir Med1998; 92:659-663.
  50. Christopher DJ, Peter JV, Cherian AM: Blind pleural biopsy using a Tru-cut needle in moderate to large pleural effusion—an experience.  Singapore Med J1998; 39:196-199.
  51. Weissberg D, Schachner A: Video-assisted thoracic surgery—state of the art.  Ann Ital Chir2000; 71:539-543.
  52. Martensson G, Pettersson K, Thiringer G: Differentiation between malignant and nonmalignant pleural effusion.  Eur J Respir Dis1985; 67:326-334.
  53. Ernst A, Hersh CP, Herth F, et al: A novel instrument for the evaluation of the pleural space: an experience in 34 patients.  Chest2002; 122:1530-1534.
  54. Roth BJ, Johnson DH, Einhorn LH, et al: Randomized study of cyclophosphamide, doxorubicin, and vincristine versus etoposide and cisplatin versus alternation of these two regimens in extensive small-cell lung cancer: a phase III trial of the Southeastern Cancer Study Group.  J Clin Oncol1992; 10:282-291.
  55. Noda K, Nishiwaki Y, Kawahara M, et al: Irinotecan plus cisplatin compared with etoposide plus cisplatin for extensive small-cell lung cancer.  N Engl J Med2002; 346:85-91.
  56. Bosl GJ, Motzer RJ: Testicular germ-cell cancer.  N Engl J Med1997; 337:242-253.
  57. Chabner BA, Stoller RG, Hande K, et al: Methotrexate disposition in humans: case studies in ovarian cancer and following high-dose infusion.  Drug Metab Rev1978; 8:107-117.
  58. Kinasewitz G, Fishman N: Pleural effusions.   In: Fishman N, ed. Pulmonary Diseases and Disorders,  New York: McGraw-Hill; 1988:750-801.
  59. Gross NJ: Pulmonary effects of radiation therapy.  Ann Intern Med1977; 86:81-92.
  60. Bouros D, Froudarakis M, Siafakas NM: Pleurodesis: everything flows.  Chest2000; 118:577-579.
  61. Rodriguez-Panadero F, Antony VB: Pleurodesis: state of the art.  Eur Respir J1997; 10:1648-1654.
  62. Kroegel C, Antony VB: Immunobiology of pleural inflammation: potential implications for pathogenesis, diagnosis and therapy.  Eur Respir J1997; 10:2411-2418.
  63. Antony VB, Rothfuss KJ, Godbey SW, et al: Mechanism of tetracycline-hydrochloride-induced pleurodesis. Tetracycline-hydrochloride-stimulated mesothelial cells produce a growth-factor-like activity for fibroblasts.  Am Rev Respir Dis1992; 146:1009-1013.
  64. Antunes G, Neville E: Management of malignant pleural effusions.  Thorax2000; 55:981-983.
  65. Clementsen P, Evald T, Grode G, et al: Treatment of malignant pleural effusion: pleurodesis using a small percutaneous catheter. A prospective randomized study.  Respir Med1998; 92:593-596.
  66. Saffran L, Ost DE, Fein AM, Schiff MJ: Outpatient pleurodesis of malignant pleural effusions using a small-bore pigtail catheter.  Chest2000; 118:417-421.
  67. Heffner JE, Unruh LC: Tetracycline pleurodesis. Adios, farewell, adieu.  Chest1992; 101:5-7.
  68. Costabel U: Adieu, tetracycline pleurodesis (but not in Germany).  Chest1993; 103:984.
  69. Heffner JE, Standerfer RJ, Torstveit J, Unruh L: Clinical efficacy of doxycycline for pleurodesis.  Chest1994; 105:1743-1747.
  70. Pulsiripunya C, Youngchaiyud P, Pushpakom R, et al: The efficacy of doxycycline as a pleural sclerosing agent in malignant pleural effusion: a prospective study.  Respirology1996; 1:69-72.
  71. Kennedy L, Vaughan LM, Steed LL, Sahn SA: Sterilization of talc for pleurodesis. Available techniques, efficacy, and cost analysis.  Chest1995; 107:1032-1034.
  72. Kennedy L, Sahn SA: Talc pleurodesis for the treatment of pneumothorax and pleural effusion.  Chest1994; 106:1215-1222.
  73. Xie C, Teixeira LR, McGovern JP, Light RW: Systemic corticosteroids decrease the effectiveness of talc pleurodesis.  Am J Respir Crit Care Med1998; 157:1441-1444.
  74. Kennedy L, Rusch VW, Strange C, et al: Pleurodesis using talc slurry.  Chest1994; 106:342-346.
  75. de Campos JR, Vargas FS, de Campos WE, et al: Thoracoscopy talc poudrage: a 15-year experience.  Chest2001; 119:801-806.
  76. Campos JR, Werebe EC, Vargas FS, et al: Respiratory failure due to insufflated talc.  Lancet1997; 349:251-252.
  77. Rinaldo JE, Owens GR, Rogers RM: Adult respiratory distress syndrome following intrapleural instillation of talc.  J Thorac Cardiovasc Surg1983; 85:523-526.
  78. Light RW: Talc should not be used for pleurodesis.  Am J Respir Crit Care Med2000; 162:2024-2026.
  79. Light RW: Diseases of the pleura: the use of talc for pleurodesis.  Curr Opin Pulm Med2000; 6:255-258.
  80. Rehse DH, Aye RW, Florence MG: Respiratory failure following talc pleurodesis.  Am J Surg1999; 177:437-440.
  81. Sahn SA: Talc should be used for pleurodesis.  Am J Respir Crit Care Med2000; 162:2023-2026.
  82. Ostrowski MJ: Intracavitary therapy with bleomycin for the treatment of malignant pleural effusions.  J Surg Oncol Suppl1989; 1:7-13.
  83. Noppen M, Degreve J, Mignolet M, Vincken W: A prospective, randomised study comparing the efficacy of talc slurry and bleomycin in the treatment of malignant pleural effusions.  Acta Clin Belg1997; 52:258-262.
  84. Patz Jr EF, McAdams HP, Erasmus JJ, et al: Sclerotherapy for malignant pleural effusions: a prospective randomized trial of bleomycin vs doxycycline with small-bore catheter drainage.  Chest1998; 113:1305-1311.
  85. Walker-Renard PB, Vaughan LM, Sahn SA: Chemical pleurodesis for malignant pleural effusions.  Ann Intern Med1994; 120:56-64.
  86. Zimmer PW, Hill M, Casey K, Harvey E, Low DE: Prospective randomized trial of talc slurry vs bleomycin in pleurodesis for symptomatic malignant pleural effusions.  Chest1997; 112:430-434.
  87. Groth G, Gatzemeier U, Haussingen K, et al: Intrapleural palliative treatment of malignant pleural effusions with mitoxantrone versus placebo (pleural tube alone).  Ann Oncol1991; 2:213-215.
  88. Sorensen PG, Svendsen TL, Enk B: Treatment of malignant pleural effusion with drainage, with and without instillation of talc.  Eur J Respir Dis1984; 65:131-135.
  89. Bertrand PC, Regnard JF, Spaggiari L, et al: Immediate and long-term results after surgical treatment of primary spontaneous pneumothorax by VATS.  Ann Thorac Surg1996; 61:1641-1645.
  90. Heffner JE, Nietert PJ, Barbieri C: Pleural fluid pH as a predictor of pleurodesis failure: analysis of primary data.  Chest2000; 117:87-95.
  91. American Thoracic Society : Management of malignant pleural effusions.  Am J Respir Crit Care Med2000; 162:1987-2001.
  92. Petrou M, Kaplan D, Goldstraw P: Management of recurrent malignant pleural effusions. The complementary role of talc pleurodesis and pleuroperitoneal shunting.  Cancer1995; 75:801-805.
  93. al Kattan KM, Kaplan DK, Goldstraw P: The non-functioning pleuroperitoneal shunt: revise or replace?.  Thorac Cardiovasc Surg1994; 42:310-312.
  94. Shoji T, Tanaka F, Yanagihara K, et al: Phase II study of repeated intrapleural chemotherapy using implantable access system for management of malignant pleural effusion.  Chest2002; 121:821-824.
  95. Driesen P, Boutin C, Viallat JR, et al: Implantable access system for prolonged intrapleural immunotherapy.  Eur Respir J1994; 7:1889-1892.
  96. Pollak JS: Malignant pleural effusions: treatment with tunneled long-term drainage catheters.  Curr Opin Pulm Med2002; 8:302-307.
  97. Putnam Jr JB, Light RW, Rodriguez RM, et al: A randomized comparison of indwelling pleural catheter and doxycycline pleurodesis in the management of malignant pleural effusions.  Cancer1999; 86:1992-1999.
  98. Hancock EW: Neoplastic pericardial disease.  Cardiol Clin1990; 8:673-682.
  99. Ewart W: Practical aids in the diagnosis of pericardial effusion, in connection with the question as to surgical treatment.  Br Med J1896; 1:717-721.
  100. Wilkes JD, Fidias P, Vaickus L, Perez RP: Malignancy-related pericardial effusion. 127 cases from the Roswell Park Cancer Institute.  Cancer1995; 76:1377-1387.
  101. Abraham KP, Reddy V, Gattuso P: Neoplasms metastatic to the heart: review of 3314 consecutive autopsies.  Am J Cardiovasc Pathol1990; 3:195-198.
  102. Porte HL, Janecki-Delebecq TJ, Finzi L, et al: Pericardoscopy for primary management of pericardial effusion in cancer patients.  Eur J Cardiothorac Surg1999; 16:287-291.
  103. Chong HH, Plotnick GD: Pericardial effusion and tamponade: evaluation, imaging modalities, and management.  Compr Ther1995; 21:378-385.
  104. Sechtem U, Tscholakoff D, Higgins CB: MRI of the abnormal pericardium.  Am J Roentgenol1986; 147:245-252.
  105. Kilpatrick ZM, Chapman CB: On pericardiocentesis.  Am J Cardiol1965; 16:722-728.
  106. Hall JB, Schmidt LD: Emergencies in critical care.   In: Hall JB, Schmidt GA, Wood LD, Crinc PF, ed. Principles of Critical Care,  New York: McGraw Hill; 1997:1405-1410.
  107. Gascho JA, Martins JB, Marcus ML, Kerber RE: Effects of volume expansion and vasodilators in acute pericardial tamponade.  Am J Physiol1981; 240:H49-H53.
  108. Permanyer-Miralda , Sagrista-Sauleda J, Soler-Soler J: Primary acute pericardial disease: a prospective series of 231 consecutive patients.  Am J Cardiol1985; 56:623-630.
  109. Sagrista-Sauleda J, Merce J, Permanyer-Miralda G, Soler-Soler J: Clinical clues to the causes of large pericardial effusions.  Am J Med2000; 109:95-101.
  110. Bardales RH, Stanley MW, Schaefer RF, et al: Secondary pericardial malignancies: a critical appraisal of the role of cytology, pericardial biopsy, and DNA ploidy analysis.  Am J Clin Pathol1996; 106:29-34.
  111. Monte SA, Ehya H, Lang WR: Positive effusion cytology as the initial presentation of malignancy.  Acta Cytol1987; 31:448-452.
  112. Stewart JR, Fajardo LF, Gillette SM, Constine LS: Radiation injury to the heart.  Int J Radiat Oncol Biol Phys1995; 31:1205-1211.
  113. Seferovic PM, Ristic AD, Maksimovic R, et al: Flexible percutaneous pericardioscopy: inherent drawbacks and recent advances.  Herz2000; 25:741-747.
  114. Maisch B, Ristic AD, Rupp H, Spodick DH: Pericardial access using the PerDUCER and flexible percutaneous pericardioscopy.  Am J Cardiol2001; 88:1323-1326.
  115. Vaitkus PT, Herrmann HC, LeWinter MM: Treatment of malignant pericardial effusion.  JAMA1994; 272:59-64.
  116. Strupp C, Germing U, Trommer I, et al: Pericardial effusion in chronic myelomonocytic leukemia (CMML): a case report and review of the literature.  Leuk Res2000; 24:1059-1062.
  117. Rolla G, Bertero MT, Pastena G, et al: Primary lymphoma of the heart. A case report and review of the literature.  Leuk Res2002; 26:117-120.
  118. Arya LS, Narain S, Thavaraj V, et al: Leukemic pericardial effusion causing cardiac tamponade.  Med Pediatr Oncol2002; 38:282-284.
  119. Saad MF, Frazier OH, Hickey RC, Samaan NA: Intrapericardial pheochromocytoma.  Am J Med1983; 75:371-376.
  120. Tsang TS, Seward JB, Barnes ME, et al: Outcomes of primary and secondary treatment of pericardial effusion in patients with malignancy.  Mayo Clin Proc2000; 75:248-253.
  121. Larrey EL: New surgical procedure to open the pericardium in the case of fluid in the cavity.  Clin Chir1829; 36:303-337.
  122. Okamoto H, Shinkai T, Yamakido M, Saijo N: Cardiac tamponade caused by primary lung cancer and the management of pericardial effusion.  Cancer1993; 71:93-98.
  123. Allen KB, Faber LP, Warren WH, Shaar CJ: Pericardial effusion: subxiphoid pericardiostomy versus percutaneous catheter drainage.  Ann Thorac Surg1999; 67:437-440.
  124. Piehler JM, Pluth JR, Schaff HV, et al: Surgical management of effusive pericardial disease. influence of extent of pericardial resection on clinical course.  J Thorac Cardiovasc Surg1985; 90:506-516.
  125. Hazelrigg SR, Mack MJ, Landreneau RJ, et al: Thoracoscopic pericardiectomy for effusive pericardial disease.  Ann Thorac Surg1993; 56:792-795.
  126. Naunheim KS, Kesler KA, Fiore AC, et al: Pericardial drainage: subxiphoid vs. transthoracic approach.  Eur J Cardiothorac Surg1991; 5:99-104.
  127. Park JS, Rentschler R, Wilbur D: Surgical management of pericardial effusion in patients with malignancies. Comparison of subxiphoid window versus pericardiectomy.  Cancer1991; 67:76-80.
  128. Wang N, Feikes JR, Mogensen T, et al: Pericardio-peritoneal shunt: an alternative treatment for malignant pericardial effusion.  Ann Thorac Surg1994; 57:289-292.
  129. Ziskind AA, Pearce AC, Lemmon CC, et al: Percutaneous balloon pericardiotomy for the treatment of cardiac tamponade and large pericardial effusions: description of technique and report of the first 50 cases.  J Am Coll Cardiol1993; 21:1-5.
  130. Galli M, Politi A, Pedretti F, et al: Percutaneous balloon pericardiotomy for malignant pericardial tamponade.  Chest1995; 108:1499-1501.
  131. Bertrand O, Legrand V, Kulbertus H: Percutaneous balloon pericardiotomy: a case report and analysis of mechanism of action.  Cathet Cardiovasc Diagn1996; 38:180-182.
  132. Fiocco M, Krasna MJ: The management of malignant pleural and pericardial effusions.  Hematol Oncol Clin North Am1997; 11:253-265.
  133. Maher EA, Shepherd FA, Todd TJ: Pericardial sclerosis as the primary management of malignant pericardial effusion and cardiac tamponade.  J Thorac Cardiovasc Surg1996; 112:637-643.
  134. Shepherd FA, Morgan C, Evans WK, et al: Medical management of malignant pericardial effusion by tetracycline sclerosis.  Am J Cardiol1987; 60:1161-1166.
  135. Liu G, Crump M, Goss PE, et al: Prospective comparison of the sclerosing agents doxycycline and bleomycin for the primary management of malignant pericardial effusion and cardiac tamponade.  J Clin Oncol1996; 14:3141-3147.
  136. Imamura T, Tamura K, Takenaga M, et al: Intrapericardial OK-432 instillation for the management of malignant pericardial effusion.  Cancer1991; 68:259-263.
  137. Moriya T, Takiguchi Y, Tabeta H, et al: Controlling malignant pericardial effusion by intrapericardial carboplatin administration in patients with primary non-small-cell lung cancer.  Br J Cancer2000; 83:858-862.
  138. Fiorentino MV, Daniele O, Morandi P, et al: Intrapericardial instillation of platin in malignant pericardial effusion.  Cancer1988; 62:1904-1906.
  139. Kawashima O, Kurihara T, Kamiyoshihara M, et al: Management of malignant pericardial effusion resulting from recurrent cancer with local instillation of aclarubicin hydrochloride.  Am J Clin Oncol1999; 22:396-398.
  140. Kuhn K, Purea H, Selbach J, Westerhausen M: Treatment with locally applied mitoxantrone.  Acta Med Austriaca1989; 16:87-90.
  141. Wilkins III HE, Cacioppo J, Connolly MM, et al: Intrapericardial interferon in the management of malignant pericardial effusion.  Chest1998; 114:330-331.
  142. Laham RJ, Cohen DJ, Kuntz RE, et al: Pericardial effusion in patients with cancer: outcome with contemporary management strategies.  Heart1996; 75:67-71.
  143. Runyon BA: Care of patients with ascites.  N Engl J Med1994; 330:337-342.
  144. Lifshitz S: Ascites, pathophysiology and control measures.  Int J Radiat Oncol Biol Phys1982; 8:1423-1426.
  145. Hirabayashi K, Graham J: Genesis of ascites in ovarian cancer.  Am J Obstet Gynecol1970; 106:492-497.
  146. Coates G, Bush RS, Aspin N: A study of ascites using lymphoscintigraphy with 99mTc-sulfur colloid.  Radiology1973; 107:577-583.
  147. Bronskill MJ, Bush RS, Ege GN: A quantitative measurement of peritoneal drainage in malignant ascites.  Cancer1977; 40:2375-2380.
  148. Yamamoto S, Konishi I, Mandai M, et al: Expression of vascular endothelial growth factor (VEGF) in epithelial ovarian neoplasms: correlation with clinicopathology and patient survival, and analysis of serum VEGF levels.  Br J Cancer1997; 76:1221-1227.
  149. Zebrowski BK, Liu , Ramirez K, Akagi Y, et al: Markedly elevated levels of vascular endothelial growth factor in malignant ascites.  Ann Surg Oncol1999; 6:373-378.
  150. Mesiano S, Ferrara N, Jaffe RB: Role of vascular endothelial growth factor in ovarian cancer: inhibition of ascites formation by immunoneutralization.  Am J Pathol1998; 153:1249-1256.
  151. Bertoglio S, Melioli G, Baldini E, et al: Intraperitoneal infusion of recombinant interleukin-2 in malignant ascites in patients with gastrointestinal and ovarian cancer.  Acta Med Austriaca1989; 16:81-83.
  152. Ott MG, Mannel DN, Gallati H, et al: Peripheral natural killer cell activity and intraperitoneal soluble p55 tumor necrosis factor receptor in patients with malignant ascites: two possible indicators for response to intraperitoneal combined tumor necrosis factor alpha and interferon gamma treatment.  Cancer Immunol Immunother1996; 42:31-37.
  153. Parsons SL, Watson SA, Steele RJ: Phase I/II trial of batimastat, a matrix metalloproteinase inhibitor, in patients with malignant ascites.  Eur J Surg Oncol1997; 23:526-531.
  154. Beatty JD, Romero C, Brown PW, et al: Clinical value of carcinoembryonic antigen: diagnosis, prognosis, and follow-up of patients with cancer.  Arch Surg1979; 114:563-567.
  155. Tamsma JT, Keizer HJ, Meinders AE: Pathogenesis of malignant ascites: Starling's law of capillary hemodynamics revisited.  Ann Oncol2001; 12:1353-1357.
  156. Cattau Jr EL, Benjamin SB, Knuff TE, Castell DO: The accuracy of the physical examination in the diagnosis of suspected ascites.  JAMA1982; 247:1164-1166.
  157. Inadomi J, Cello JP, Koch J: Ultrasonographic determination of ascitic volume.  Hepatology1996; 24:549-551.
  158. Runyon BA: Management of adult patients with ascites caused by cirrhosis.  Hepatology1998; 27:264-272.
  159. Press OW, Press NO, Kaufman SD: Evaluation and management of chylous ascites.  Ann Intern Med1982; 96:358-364.
  160. Runyon BA, Montano AA, Akriviadis EA, et al: The serum-ascites albumin gradient is superior to the exudate-transudate concept in the differential diagnosis of ascites.  Ann Intern Med1992; 117:215-220.
  161. Akriviadis EA: Hemoperitoneum in patients with ascites.  Am J Gastroenterol1997; 92:567-575.
  162. Runyon BA: Low-protein-concentration ascitic fluid is predisposed to spontaneous bacterial peritonitis.  Gastroenterology1986; 91:1343-1346.
  163. Siddiqui RA, Kochhar R, Singh V, et al: Evaluation of fibronectin as a marker of malignant ascites.  J Gastroenterol Hepatol1992; 7:161-164.
  164. Runyon BA: Malignancy-related ascites and ascitic fluid “humoral tests of malignancy.”.  J Clin Gastroenterol1994; 18:94-98.
  165. Cardozo PL: A critical evaluation of 3000 cytologic cases of pleural fluid, ascitic fluid, and pericardial fluid.  Acta Cytol1960; 10:455.
  166. Loewenstein MS, Rittgers RA, Feinerman AE, et al: Carcinoembryonic antigen assay of ascites and detection of malignancy.  Ann Intern Med1978; 88:635-638.
  167. Gershenson DM, Silva EG, Tortolero-Luna G, et al: Serous borderline tumors of the ovary with noninvasive peritoneal implants.  Cancer1998; 83:2157-2163.
  168. Gershenson DM, Silva EG, Levy L, et al: Ovarian serous borderline tumors with invasive peritoneal implants.  Cancer1998; 82:1096-1103.
  169. Strnad CM, Grosh WW, Baxter J, et al: Peritoneal carcinomatosis of unknown primary site in women. A distinctive subset of adenocarcinoma.  Ann Intern Med1989; 111:213-217.
  170. van Dam PA, DeCloedt J, Tjalma WA, et al: Trocar implantation metastasis after laparoscopy in patients with advanced ovarian cancer: can the risk be reduced?.  Am J Obstet Gynecol1999; 181:536-541.
  171. Menzies RI, Fitzgerald JM, Mulpeter K: Laparoscopic diagnosis of ascites in Lesotho.  Br Med J1985; 291:473-475.
  172. Osterlee J: Peritoneovenous shunting for ascites in cancer patients.  Br J Surg1980; 67:145-160.
  173. Greenway B, Johnson PJ, Williams R: Control of malignant ascites with spironolactone.  Br J Surg1982; 69:441-442.
  174. Razis DV, Athanasiou A, Dadiotis L: Diuretics in malignant effusions and edemas of generalized cancer.  J Med1976; 7:449-461.
  175. Lee CW, Bociek G, Faught W: A survey of practice in management of malignant ascites.  J Pain Symptom Manage1998; 16:96-101.
  176. Pockros PJ, Esrason KT, Nguyen C, et al: Mobilization of malignant ascites with diuretics is dependent on ascitic fluid characteristics.  Gastroenterology1992; 103:1302-1306.
  177. Parsons SL, Lang MW, Steele RJ: Malignant ascites: a 2-year review from a teaching hospital.  Eur J Surg Oncol1996; 22:237-239.
  178. Gines P, Tito L, Arroyo V, et al: Randomized comparative study of therapeutic paracentesis with and without intravenous albumin in cirrhosis.  Gastroenterology1988; 94:1493-1502.
  179. Straus AK, Roseman DL, Shapiro TM: Peritoneovenous shunting in the management of malignant ascites.  Arch Surg1979; 114:489-491.
  180. Edney JA, Hill A, Armstrong D: Peritoneovenous shunts palliate malignant ascites.  Am J Surg1989; 158:598-601.
  181. Gough IR, Balderson GA: Malignant ascites. A comparison of peritoneovenous shunting and nonoperative management.  Cancer1993; 71:2377-2382.
  182. Schumacher DL, Saclarides TJ, Staren ED: Peritoneovenous shunts for palliation of the patient with malignant ascites.  Ann Surg Oncol1994; 1:378-381.
  183. Souter RG, Wells C, Tarin D, Kettlewell MG: Surgical and pathologic complications associated with peritoneovenous shunts in management of malignant ascites.  Cancer1985; 55:1973-1978.
  184. Ragni MV, Lewis JH, Spero JA: Ascites-induced LeVeen shunt coagulopathy.  Ann Surg1983; 198:91-95.
  185. Scott-Coombes DM, Whawell SA, Vipond MN, et al: Fibrinolytic activity of ascites caused by alcoholic cirrhosis and peritoneal malignancy.  Gut1993; 34:1120-1122.
  186. Gleysteen JJ, Hussey CV, Heckman MG: The cause of coagulopathy after peritoneovenous shunt for malignant ascites.  Arch Surg1990; 125:474-477.
  187. Smith RR, Sternberg SS, Paglia MA, Golbey RB: Fatal pulmonary tumor embolization following peritoneovenous shunting for malignant ascites.  J Surg Oncol1981; 16:27-35.
  188. Borger JA, Pitel P, Crump G: Management of malignant ascites with a vascular port.  J Pediatr Surg1993; 28:1605-1606.
  189. Richard III HM, Coldwell DM, Boyd-Kranis RL, et al: Pleurx tunneled catheter in the management of malignant ascites.  J Vasc Interv Radiol2001; 12:373-375.
  190. Sabatelli FW, Glassman ML, Kerns SR, Hawkins Jr IF: Permanent indwelling peritoneal access device for the management of malignant ascites.  Cardiovasc Intervent Radiol1994; 17:292-294.
  191. Lomas DA, Wallis PJ, Stockley RA: Palliation of malignant ascites with a Tenckhoff catheter.  Thorax1989; 44:828.
  192. Belfort MA, Stevens PJ, DeHaek K, et al: A new approach to the management of malignant ascites; a permanently implanted abdominal drain.  Eur J Surg Oncol1990; 16:47-53.
  193. Flowers III RH, Schwenzer KJ, Kopel RF, et al: Efficacy of an attachable subcutaneous cuff for the prevention of intravascular catheter-related infection. A randomized, controlled trial.  JAMA1989; 261:878-883.
  194. Sugarbaker PH: Peritonectomy procedures.  Ann Surg1995; 221:29-42.
  195. Sugarbaker PH, Jablonski KA: Prognostic features of 51 colorectal and 130 appendiceal cancer patients with peritoneal carcinomatosis treated by cytoreductive surgery and intraperitoneal chemotherapy.  Ann Surg1995; 221:124-132.
  196. Sugarbaker PH: Intraperitoneal chemotherapy for treatment and prevention of peritoneal carcinomatosis and sarcomatosis.  Dis Colon Rectum1994; 37:S115-S122.
  197. Dedrick RL, Myers CE, Bungay PM, DeVita Jr VT: Pharmacokinetic rationale for peritoneal drug administration in the treatment of ovarian cancer.  Cancer Treat Rep1978; 62:1-11.
  198. Alberts DS, Liu PY, Hannigan EV, et al: Intraperitoneal cisplatin plus intravenous cyclophosphamide versus intravenous cisplatin plus intravenous cyclophosphamide for stage III ovarian cancer.  N Engl J Med1996; 335:1950-1955.