Mechanisms of pleural liquid accumulation in disease

INTRODUCTION — 

In the normal pleural space, there is a steady state in which there is a roughly equal rate of the formation (entry) and absorption (exit) of liquid (figure 1). This balance must be disturbed to produce a pleural effusion [1,2]. This topic discusses the mechanisms involved in pleural fluid accumulation in pathologic states. Mechanisms that underlie turnover in the normal state are discussed separately. (See "Mechanisms of pleural liquid turnover in the normal state".)

BACKGROUND — 

To generate a pleural effusion, there must be an increase in entry rate and/or a reduction in exit rate. To generate a clinically relevant effusion, it is likely that both mechanisms contribute to effusion formation for the following reasons:

●An isolated increase in entry rate, unless large and sustained, is unlikely to cause a clinically significant effusion because the absorbing pleural lymphatics have a large reserve capacity to deal with excess pleural liquid. If, for example, artificial hydrothoraces are instilled into the pleural space of awake sheep, the exit rate can increase to 0.28 mL/kg per hour, which is 28 times the baseline rate [3].

●An isolated decrease in exit rate is also unlikely to cause a large effusion because the normal entry rate is low. Even if the exit of liquid ceased entirely, accumulation of liquid would take many days to become evident. As an example, the normal entry rate of 0.01 mL/kg per hour is equivalent to a total of 12 mL/day in a 50 kg female; at this rate of entry, without any exit of liquid, it would take more than one month for 500 mL to accumulate in the pleural space. Of note, the effusion would presumably be a transudate because the normal liquid entering the pleural space is low in protein.

INCREASED FLUID ENTRY — 

Factors involved in common clinical scenarios can increase fluid flow into the pleural space. Such factors, described in this section, can include increased venous pressure, decreased pleural pressure, or increased vascular permeability due to inflammation, infection, or malignancy. Normally, liquid filters out of systemic microvessels based on a balance of hydrostatic and osmotic forces across a semipermeable membrane [4,5]. These forces are described in the Starling equation, in which the hydrostatic forces that force water out of the vessel are balanced by osmotic forces that reabsorb water back into the vessel [6,7].

In this equation, k is liquid conductance of the microvascular barrier, Pmv and Ppmv represent hydrostatic pressure in the microvascular and perimicrovascular compartments (respectively), is the reflection coefficient for total protein and ranges from 0 (completely permeable) to 1 (completely impermeable), and πmv and πpmv represent protein osmotic pressure in microvascular and perimicrovascular liquids, respectively. In normal microvessels, there is ongoing filtration of a small amount of low-protein liquid. The flow from the microvessels can increase with changes in various parameters of the Starling equation.

Increase in permeability — Increases in permeability of blood vessels result from inflammation, infection, or malignancy, all common causes of pleural effusions. If the endothelial barrier becomes more permeable to liquid and protein, for example, there will be an increase in flow of a higher-protein liquid. Because absorption does not alter the protein concentration of pleural liquid, pleural liquid with a high protein concentration indicates its origin from a circulation with increased permeability. (See "Mechanisms of pleural liquid turnover in the normal state", section on 'Evidence for bulk flow'.)

Increase in microvascular pressure — As mentioned above, an elevation in microvascular pressure (Pmv) is usually caused by an elevation in venous pressure. Increases in arterial pressure are less likely to be transmitted to the microvessels because of the high precapillary resistance and autoregulation of arteriolar tone. Either systemic or pulmonary venous pressure can lead to increased filtration and increased entry of fluid into the pleural space. Clinical conditions with elevated systemic venous pressure include congestive heart failure (CHF) with right heart failure, pulmonary embolism, and central venous obstruction. The major clinical condition associated with increased pulmonary venous pressure is CHF.

Elevations in systemic venous pressure alone (which would mostly affect the parietal pleura) would be associated with a greater sieving of proteins, leading to a filtrate with a lower protein concentration than normal (with a pleural liquid-to-serum protein ratio of <0.15). Of course, most transudative effusions have a pleural liquid-to-serum protein ratio much higher than this, between 0.4 and 0.5. This fact demonstrates that most liquid must arise from a source other than the systemic circulation of the pleural membranes. The likely source is the large nonsystemic circulation adjacent to the pleural space, namely the pulmonary circulation of the nearby lung. In the normal state, lung interstitial liquid (eg, lymph) filtered from the low-pressure pulmonary circulation has a protein concentration ratio (lung-to-serum protein concentration ratio) of 0.7, but this ratio falls to 0.4 to 0.5 with increased flow due to increased pulmonary microvascular pressures [8]. As shown in studies of anesthetized sheep, this lung interstitial edema liquid is the likely source of the majority of the hydrostatic pleural effusion [9].

Lung interstitial edema liquid reaches the pleural space in a complex way. When the rate of filtrate formation exceeds the absorptive capacity of the lung lymphatics, the filtrate accumulates in the peribronchovascular spaces ("cuffs") [10]. Once in these interstitial spaces, the liquid is not accessible to lung lymphatics [11]. Thus, although the lymphatics are important in removing liquid as it is filtered from the pulmonary circulation, they cannot account for the clearance of edema already established in the lung [12]. This interstitial edema probably leaves the lung by flowing in response to pressure gradients along the interstitial spaces (interlobular septae, peribronchovascular bundles, and visceral pleura) of the lung toward either the mediastinum or the pleural space. The entry of large amounts of lung interstitial liquid into the pleural space will elevate the overall protein concentration of the pleural liquid, giving a ratio of 0.4 to 0.5, the expected range for a transudative effusion [9].

The vast majority of transudative effusions are caused by pulmonary venous pressure elevation due to CHF and have protein ratios of approximately 0.4. If a transudative effusion has a lower protein concentration, it may be caused by systemic venous pressure elevation, such as by central venous obstruction.

Decrease in pleural pressure — Decreased pleural pressure is the likely explanation for the formation of pleural effusions in the setting of atelectasis.

A decrease in pleural pressure, as seen with significant atelectasis, alters the balance of forces described in the Starling equation by reducing the pressures surrounding the adjacent microvessels. This decrease in perimicrovascular pressures (Ppmv) enhances filtration across the microvascular barrier. Thus, fluid will accumulate until the pressure balance is restored.

Decrease in plasma osmotic pressure — While a low albumin or protein concentration may be an infrequent cause of effusions by itself, it may contribute to effusion formation by other causes [13].

Hypoproteinemia (due to hypoalbuminemia) will decrease the plasma oncotic pressure (πmv), thereby increasing the forces favoring filtration until the balance is restored. By itself, hypoproteinemia can probably induce small effusions with a low protein concentration. In addition, hypoproteinemia can lower the threshold for effusion formation when other Starling forces are changed. In a study of hospitalized patients with acquired immunodeficiency syndrome (AIDS), for example, hypoproteinemia alone was the apparent cause of 19 percent of all pleural effusions [14]. However, together with other factors, a lower serum protein concentration may have contributed to effusion formation in many more patients because, in general, the patients with effusions had a lower serum albumin concentration than those without effusion (2.5 versus 3.4 g/dL). In another study, hypoalbuminemia may have enhanced the size of pleural effusions, perhaps by shifting the balance toward increased filtration; in children with parapneumonic pleural effusions, those with large effusions had a lower mean serum albumin concentration than those with small effusions (2.7 versus 3.7 g/dL) [15]. In another study in medical intensive care unit patients, those with effusions had a lower serum albumin concentration than those without effusions (2.4 versus 3 g/dL) [16].

DECREASED FLUID EXIT — 

A decrease in exit rate indicates a reduction in clearance of fluid from the pleural space, a clearance that largely takes place via the parietal pleural lymphatics. Because lymphatic function is poorly understood, much of this discussion is speculative. Unlike blood vessels, lymphatic vessels have one-way valves. Lymph is propelled actively by rhythmic muscular contractions of lymphatic smooth muscle and passively by compressions from the respiratory motions of the chest wall. In addition, flow is affected by lymphatic patency, availability of liquid at the lymphatic orifices in the parietal pleura, and the pressures influencing filling (pleural pressure) and emptying (systemic venous pressure) of lymphatics [17,18].

Intrinsic factors — A number of factors may inhibit the ability of lymphatics to contract, including:

●Cytokines and products of inflammation (eg, endotoxins)

●Endocrine abnormalities (eg, hypothyroidism)

●Injury due to radiation or drugs (eg, chemotherapeutic agents)

●Infiltration of lymphatics by cancer

●Anatomic abnormalities (eg, yellow nail syndrome)

Extrinsic factors — Multiple extrinsic factors may inhibit lymphatic function even when the lymphatics themselves are normal. These include:

●Limitation of respiratory motion (eg, diaphragm paralysis, lung collapse, pneumothorax)

●Mechanical compression of lymphatics (eg, pleural fibrosis, pleural granulomas)

●Blockage of lymphatic stomata (eg, fibrin deposition on pleural surface, pleural malignancy)

●Decreased intrapleural pressure (eg, atelectasis)

●Increased systemic venous pressure – Acutely, increases in venous pressure may decrease lymphatic flow because of the higher downstream pressures into which the lymphatics empty; chronically, the lymphatics may be able to adapt.

●Decreased availability of fluid at the lymphatics – After pneumothorax, for example, fluid may not remain in contact with lymphatic stomata and may accumulate at the dependent portion of the pleural space as a small effusion.

INTERPLAY OF PATHOGENETIC MECHANISMS IN DISEASE STATES

One disease or a combination of diseases — As discussed above, no single mechanism is likely to explain the development of a pleural effusion. It is likely that, for many effusions, multiple factors contribute to effusion formation. In addition, an alteration of one mechanism can lower the threshold for effusion formation later by another mechanism. As an example, a decreased exit rate due to malignant infiltration or fibrosis of lymphatics can gradually reduce the capacity of the lymphatics to drain pleural liquid. As long as the entry rate remains low, an effusion will not form; however, if the entry rate should increase, the limitation of the exit rate would reduce the ability of the lymphatics to handle the increased fluid and an effusion could then accumulate. Thus, it has been proposed that a single disease can alter both the entry and exit rates and produce a clinically relevant effusion or more than one disease can cooperate to form an effusion, perhaps with one condition increasing entry while the other decreases the exit [1].

There are some data on the entry and exit rates of pleural liquid in patients with pleural effusions. In two early studies of the turnover of pleural liquid in effusions of patients with a variety of different diseases, the entry rates were highest in tuberculosis but similar among the other disorders [19,20]. The exit rates were low in malignant and tuberculous effusions compared with the rates in effusions due to cardiac failure and pulmonary embolism (PE). In one study, for example, calculated lymphatic flow for patients with PE was 0.18 mL/kg per hour, which is similar to the maximal exit rate measured for sheep (0.28 mL/kg per hour); in comparison, lymphatic flow was lower in the patients with carcinoma or tuberculosis (0.06 and 0.08 mL/kg per hour, respectively), suggesting that a reduced lymphatic capacity contributed to the formation of these effusions [20]. Interestingly, in patients with tuberculous effusions, the entry rate decreased and the exit rate increased after treatment with prednisone.

Multiple causes have also been implicated in the formation of clinical pleural effusions and causes of an effusion can change over time [21]. In one study of both exudative and transudative bilateral effusions, 47 and 83 percent (respectively) were thought to have multiple causes, including congestive heart failure (CHF), kidney failure, hypoalbuminemia, and cancer [22]. In a prospective observational study of 130 consecutive patients with unilateral pleural effusions, 30 percent were determined to have more than one underlying cause for the effusion [23]. Often this additional cause was CHF but sometimes also pleural infection or malignancy.

Disease of the pleural membranes — Direct involvement of pleural membranes by disease can lead to a pleural effusion by increasing the formation of liquid and interfering with parietal pleural lymphatic function (figure 2). For patients with pleural malignancy or tuberculosis, the synergistic combination of increased entry and decreased exit of liquid may explain the massive effusions that can accumulate.

Malignancy — Malignant effusions are probably caused by both mechanisms, increased entry plus decreased exit of fluid. There are patients with rapid entry rates, which can be recognized clinically because the effusion accumulates rapidly after drainage or has a high chest tube drainage rate. In this setting, the tumor has presumably extensively infiltrated the pleural membranes, leading to increased filtration, or is producing cytokines, such as vascular endothelial growth factor (VEGF), that increase capillary permeability [24,25]. One study indicated that the mast cell may be the key cell producing cytokines, such as tryptase AB1 and IL1 beta, leading to increased permeability [26]. Decreased plasma osmotic pressure or decreased pleural pressure could contribute to the enhanced entry of liquid.

On the other hand, malignancy may lead to effusion formation by infiltrating the draining lymphatics or lymph nodes, thereby decreasing the exit rate. In some cases of lymphatic involvement, the decrease in the exit rate appears to be an important mechanism of effusion formation because the effusions can resolve after mediastinal irradiation of involved lymph nodes. In certain malignant effusions, extrapleural involvement of draining lymphatics may be the sole mechanism of effusion formation. Such an isolated exit block may explain the existence of transudative effusions, which have been described in approximately 10 percent of patients with malignant effusions [27].

Tuberculosis — In tuberculous pleural effusions, the pleural membranes are infiltrated with granulomas, an infiltration that may lead both to an increased entry rate and to a decreased exit rate of liquid. The high protein concentration of tuberculous pleural effusions indicates its origin from microvessels altered by an intense inflammation [20]. The exit rate, on the other hand, has been found to be low in both patients and animals with experimental models of Bacillus Calmette-Guerin vaccine-induced pleurisy [19,20,28]. The low exit rate may indicate that inflammation of the parietal pleural by tuberculosis is extensive and intense enough to compromise pleural lymphatic function.

Pleural microvessel hydrostatic pressure elevation — An increase in the microvascular pressure of pleural vessels can increase filtration in a variety of ways. Systemic venous pressure elevation both increases filtration from the parietal pleural microvessels and decreases lymphatic drainage into the venous system (figure 3). In comparison, an elevation in pulmonary venous pressure increases filtration into the lung and movement across the visceral pleura into the pleural space.

Pulmonary embolism — PE can increase entry rates of liquid by injuring pulmonary and adjacent pleural systemic circulations, by elevating hydrostatic pressures in pulmonary veins and/or systemic veins, and perhaps by lowering pleural pressure due to atelectasis. PE may also decrease exit rates of pleural liquid by increasing the systemic venous pressure (thereby hindering lymphatic drainage) or perhaps by decreasing pleural pressure (thereby hindering lymphatic filling). The observation that all effusions due to PE are exudates suggests a key role for vascular injury [29]; however, hydrostatic pressure changes probably also contribute to the formation of the effusions. One retrospective study comparing patients with PE found that those with effusions were more likely to have more severe PE and pulmonary infarction than those without effusions [30]. When microvessels are injured, small changes in hydrostatic pressure can have a large effect on fluid flux.

Superior vena cava syndrome — The mechanism of effusion formation in patients with the superior vena cava syndrome has only been studied in a few cases [31]. Apparently, the entry rate of liquid into the pleural space is increased. The most detailed study has been in one patient who was shown to have a transudative effusion with flow through the thoracostomy tube of approximately 500 mL/day [32]. In another study of volume-loaded dogs, an elevation in systemic venous pressure for two hours plus a decrease in plasma osmotic pressure led to a significant increase in pleural liquid entry and formation of an effusion [33].

The exit rate may also be decreased acutely since the lymphatics must pump against a higher downstream pressure. With chronic pressure elevations, however, the lymphatics may adapt and resume a more normal capacity. In a series of 27 patients with chronic systemic venous pressure elevation due to pulmonary hypertension (PH), for example, no effusions were found on ultrasonography [34]. In another study in which consecutive patients with PH underwent right heart catheterization and thoracic ultrasound, the presence of a pleural effusion was associated with an elevated pulmonary artery wedge pressure [35]. Thus, systemic venous pressure elevation alone does not appear to be a major cause of pleural effusions.

Brachiocephalic venous obstruction — Brachiocephalic venous obstruction has been identified as a cause of persistent (sometimes intractable) transudative effusions. In particular, the venous obstruction, often developing in a patient undergoing hemodialysis, is thought to increase pleural microvascular hydrostatic pressure and thereby increase pleural liquid formation and decrease lymphatic clearance [36]. It is important to be aware of this entity because the effusion can resolve after angioplasty of the occluded vessel [37,38].

Diseases of the lung — The lung is a potentially large source of liquid immediately adjacent to the pleural space. Lung interstitial liquid can move into the pleural space along a pressure gradient and across leaky pleural membranes (figure 4) [39].

Acute lung injury — From animal studies in which injury is limited to the lung, it is clear that large amounts of lung liquid can move into the pleural space. After chemical or hyperoxia-induced lung injury in animals, for example, an increased permeability lung edema developed and was followed by high-protein pleural effusions in approximately two hours [40-42]. With hyperoxic lung injury in rats, the movement of liquid from lung to pleural space could be traced by the movement of a specific marker [43]. In sheep given intravenous oleic acid, liquid leaving the lung across the visceral pleura could be collected in a surrounding bag and quantified as representing almost 20 percent of the lung edema [42].

These observations suggest that pleural effusions should be common in patients with acute lung injury. In a radiographic study of patients with pulmonary edema, for example, pleural effusions were found to be as common in patients with acute lung injury (36 percent) as in patients with hydrostatic pulmonary edema (40 percent) [44]. Other lung injuries, such as pneumonia or PE, can also result in effusion formation due to the movement of high-protein lung interstitial liquid into the pleural space.

Hydrostatic pulmonary edema — In heart failure, in contrast to lung injury, the abnormalities are not limited to the lung; as a result, identification of the source of the pleural liquid has been more difficult. An elevation in systemic venous and pulmonary venous pressures can lead to increased filtration from both pleural membranes, decreased absorption via pleural lymphatics, and increased filtration into the lung with movement of lung edema into the pleural space.

Several studies have addressed the contribution of lung edema to transudative pleural effusions in CHF. In one experimental study in which pressures were elevated in the systemic venous, the pulmonary venous, or both circulations, the most liquid appeared after systemic pressure elevation [33]. However, this study lasted only two hours and lung edema takes at least two hours to accumulate and then to flow to the pleural space [9,40,42]. It is therefore likely that the contribution from the lung was not yet evident. The observation in a clinical study of patients with heart failure that the presence of pleural effusions by ultrasound correlated better with elevated pulmonary venous pressures than with systemic venous pressures is compatible with the importance of a lung contribution [45]. A later study of patients with PH showed that isolated increases in systemic venous pressure, at least when chronic, did not cause effusion formation [34]. A similar study of patients with PH found that the presence of pleural effusions was associated with elevations of pulmonary venous pressure [35].

The contribution of the lung was directly evaluated by studies in anesthetized sheep in which the lung was isolated in an impermeable bag [9]. Lung edema was created by volume loading to elevate the pulmonary capillary wedge pressure by 10, 20, or 30 cm H₂O. Liquid began to leak from the lung and, by two hours after stable pressure elevation, the lung liquid flow had reached a steady state. The lung liquid had the same protein concentration as lung lymph and interstitial liquid later harvested from peribronchovascular spaces ("cuffs"). The amount of liquid flowing from the lung could account for the pleural effusions found in other closed-chest volume-loaded sheep. The edema that cleared the lung into the pleural space accounted for almost 23 to 29 percent of all edema formed. Thus, the contribution of the pleural route to edema clearance appears to be similar for hydrostatic edema as it is for the increased permeability edema formed due to acute lung injury [42]. These observations suggest that the pleural route of edema clearance may be an important additional safety factor protecting against alveolar flooding.

One of the strongest arguments for the contribution of lung edema to transudative effusions in heart failure is that the protein concentration is most similar to that of lung interstitial liquid. As discussed above, liquid derived from increased filtration across systemic vessels would be expected to have a very low protein concentration with a pleura-to-serum protein concentration ratio of <0.15. However, transudative effusions have a higher protein concentration ratio (approximately 0.4 to 0.5), similar to the ratio in pulmonary filtrate (eg, in lung lymph) [9].

The protein concentration of pleural fluid and the pleura-to-serum protein concentration ratio can increase with acute diuretic therapy, so that a true transudate may yield indices suggestive of an exudate [46-48]. Although not all studies have shown a change from transudative to exudative chemistries following diuresis [49], this phenomenon must be considered when interpreting pleural chemistries in patients following a significant diuresis [50]. Caution must also be used to avoid missing a true exudate in the setting of CHF, which could indicate a coincident PE, malignancy, or infection.

Extrapleural or extrapulmonary diseases — Excess liquid from any tissue in the body may find its way to the pleural space by passive flow toward the low pressure of the pleural space. Then, once adjacent to the pleural membranes, the excess fluid could move into the pleural space, either via holes or tears in the mediastinal pleura or diaphragm or via direct flow across the permeable pleural membranes.

Mediastinal inflammation — Effusions have been described following mediastinal inflammation due to esophageal variceal sclerotherapy and esophageal perforation. Chyle can also collect in the mediastinum from a break in the thoracic duct and can then decompress by draining into the pleural space, producing a chylothorax (see "Etiology, clinical presentation, and diagnosis of chylothorax"). Fluid from a pancreatic pseudocyst can also flow via pressure gradients into the mediastinum and then decompress into the pleural space. Usually decompression of the mediastinal collection is achieved when there is flow into one pleural space and thus, these effusions tend to be unilateral.

Peritoneal pathology (via the diaphragm) — Liquid can cross the diaphragm by diffusion or by movement across macroscopic holes, either acquired or congenital. Peritoneal liquid can potentially reach the pleural space by diffusion across the two mesothelial layers, a process which would likely be slow. Peritoneal fluid can also move rapidly across the diaphragm through acquired or congenital defects in the diaphragm [51]. The defects apparently form at points of weakness in the muscular webbing of the diaphragm. Based on observations during video-assisted thoracoscopy, the defects have various morphologies from blebs to fenestrations [52]. They may be missed, even on direct inspection of the diaphragm, unless subdiaphragmatic pressure is increased to expand them. Although some have speculated that liquid might cross the diaphragm via lymphatics, there is no evidence for direct lymphatic channels connecting the peritoneal and pleural spaces across the diaphragm [53,54]. In studies in which radiocontrast dye is instilled into the peritoneal space, the dye clearly drains into lymphatics that travel to the mediastinum; typically, no dye enters the pleural space. Therefore, the only likely mechanism for rapid movement of liquid across the diaphragm is through actual defects in the diaphragm [51].

SUMMARY AND RECOMMENDATIONS

●Increased fluid entry – Most effusions are likely caused by both increased pleural fluid formation and decreased pleural fluid clearance (figure 1). (See 'Increased fluid entry' above.)

•Increased permeability – The protein concentration of pleural liquid is a clue to its formation because the protein concentration is not altered by absorption of pleural liquid via lymphatics. (See 'Increase in permeability' above.)

•Increased microvascular pressure – Elevations in either systemic venous pressure (affecting the parietal pleura) or pulmonary venous pressure (affecting the visceral pleura) can lead to an increase in pleural liquid formation and the development of a pleural effusion (figure 3). (See 'Increase in microvascular pressure' above.)

Transudative effusions from heart failure are likely caused by entry of lung edema (lung interstitial fluid) into the pleural space (figure 4). Interstitial edema probably leaves the lung by flowing down pressure gradients along the interstitial spaces of the lung (interlobular septae, peribronchovascular bundles and visceral pleura) toward either the mediastinum or the pleural space. (See 'Increase in microvascular pressure' above.)

•Decreased pleural pressure – Decreased pleural pressure is the likely explanation for the formation of pleural effusions in the setting of atelectasis.

•Decreased osmotic pressure – Hypoproteinemia (due to hypoalbuminemia) decreases the plasma oncotic pressure, thereby increasing the forces favoring filtration. By itself, hypoproteinemia may result in small effusions but, when other Starling forces are changed (eg, increased central venous pressure), hypoproteinemia lowers the threshold for effusion formation. (See 'Decrease in plasma osmotic pressure' above.)

●Decreased fluid exit – A reduction in lymphatic function will decrease the absorption rate of pleural liquid. Impairment in lymphatic function may be caused by intrinsic factors (eg, hypothyroidism, cancer infiltration, yellow nail syndrome) or extrinsic factors (eg, decreased respiratory motion from diaphragmatic paralysis, perilymphatic granulomas or cancer). (See 'Decreased fluid exit' above.)

●Mechanisms in disease states – No single mechanism likely explains the development of a pleural effusion. It is likely that, for many effusions, multiple factors contribute to effusion formation. One disease, such as malignancy or tuberculosis, may have multiple effects and by itself lead to an effusion, or often combinations of diseases cooperate to form an effusion. Thus, the clinician should be alert to more than one underlying cause of an effusion. Mechanisms that underlie specific disease states are discussed above. (See 'Interplay of pathogenetic mechanisms in disease states' above.)