Introduction
Of the approximately 1 million cases of hospitalized pneumonia each year in the United States, around 60,000 will develop frank empyema. A further 25,000 are estimated to develop pleural infection for other reasons, including trauma and iatrogenic instrumentation. The annual estimated hospital cost for treating these cases is in excess of $500 million. The morbidity and mortality rates in patients with pneumonia and associated pleural effusions are higher than for those in patients with pneumonia alone, with one study suggesting the increased relative risk for mortality was 3.4 times higher. There also appears to be a rising incidence of pleural infection internationally, but, despite continued advances in the management of this condition, morbidity and mortality have essentially remained static over the past decade.
Historical Perspective
Infection of the pleural space is an ancient disease, with the earliest recorded description more than 5000 years ago and the first consistent description of its manifestations and treatment credited to the father of modern medicine, Hippocrates. Open thoracic drainage remained the standard treatment for pleural infection until the influenza pandemic of 1919; however, there was a 70% mortality rate associated with this treatment. In 1918 the U.S. Army Empyema Commission was formed to address the problem. They noted that dogs with empyema died more often if treated with early open drainage rather than delayed intervention, and the commission recommended using the closed-tube drainage techniques described by Hewitt and Bulau. The commission’s summary recommendations were: adequate pus drainage with a closed tube, avoidance of early open drainage, obliteration of the pleural space, and proper nutritional support. A landmark paper by Graham charted the successes seen during this time, with short-term mortality plummeting to 4%. These treatment principles, described almost 100 years ago, remain essentially unchanged to this day. The discovery of penicillin in the 1940s was a major advance that led to a further reduction in mortality. Surgical techniques beyond open drainage were developed in the late 19th century, with thoracoplasty described by Estlander and Schede and with decortication described by Fowler and Beck. It was only recently, however, that video-assisted thoracic surgery (VATS) was introduced and is being increasingly used today as the operation of choice in patients requiring surgery for their pleural infection.
Incidence
There have been a number of recent reports from the United States, Canada, Europe, and Asia, all showing a dramatic increase in the incidence of pleural infection. This seems to be the case for both children and adults and, although most data are derived from developed-world populations, this pattern has been replicated around the globe. The incidence of pleural infection began to rise in the mid to late 1990s. This was captured by Grijalva and colleagues, who recently examined the trends in parapneumonic empyema in the United States over a 13-year period, demonstrating a doubling in the rates of empyema hospitalizations between 1996 and 2008, from 3.04 to 5.98 per 100,000. Similar results were demonstrated by a Canadian study, which also confirmed the significant disparity in empyema incidence between those 65 years of age and older (17 to 20 per 100,000) and those 19 years of age or younger (2 to 4 per 100,000).
Mortality rates from empyema also seem to be on the rise. A study looking at the population of Utah showed a sixfold rise in mortality from the period 1950–1975 to the period 2000–2005. In a large series, inpatients were found to have a mortality up to 18% in the short term, with those in intensive care experiencing mortality as high as 41%. In a large multicenter trial from the United Kingdom, patients with an average age of 59 were shown to have a mortality rate 1 year after treatment for empyema between 8% and 20%.
The explanation for the increase in empyema incidence is not clear. With the introduction of the heptavalent pneumococcal conjugate vaccine in 2000, a reduction in pneumococcal empyema in children from serotypes covered by the vaccine may have led to an increase of cases caused by nonvaccine serotypes. This may account for an increase in adult infection with these serotypes. This, however, does not explain the increase in staphylococcal empyema seen in the series by Grijalva and associates.
Epidemiology
Pleural infection is seen in patients of all ages but has a bimodal distribution, with a peak in childhood and a further rise in older adults. Men are affected twice as often as women, and this is illustrated in Figure 80-1 , which shows the age and sex distribution of the large cohort of adult patients from the Multicenter Intrapleural Sepsis Trial (MIST) 1 study. Pleural infection is also more common in those with diabetes, alcohol dependency, or drug addiction and those with rheumatoid arthritis. Poor dentition and aspiration have also been shown to be risk factors.
The underlying etiology of pleural infection is varied, with most cases being of community origin. Although this often results from a community-acquired pneumonia, a sizable proportion of cases show no evidence of consolidation on computed tomography (CT) imaging and are thought to have been acquired through hematologic spread or direct translocation from the oropharynx. The next largest group are hospital-acquired pleural infections, which are often the result of prolonged hospital admissions for other initial reasons, or complications following surgery or invasive procedures. For example, pleural infections and other pleural complications are common after lung transplantation. Other potential causes include direct (transdiaphragmatic) spread of abdominal sepsis, blunt or penetrating chest trauma, esophageal perforation, or rupture of a peripheral lung abscess into the pleural space.
Pathogenesis
The evolution of a pleural infection can be divided into three stages, which may overlap with each other. The first, exudative stage is characterized by the rapid outpouring of sterile pleural fluid into the pleural space. Some of this comes from the interstitial spaces of the lung and some from the parietal pleura because of increased permeability. The pleural fluid will have a low white blood cell count and lactate dehydrogenase (LDH) level, together with a normal glucose level and pH. At this stage chest tube drainage is rarely required, and antibiotics alone should suffice.
The second, fibropurulent stage evolves if bacteria invade the sterile exudative effusion. During this stage there is an accumulation of leukocytes, bacteria, and cell debris together with increased amounts of pleural fluid. Fibrin is then deposited over the visceral and parietal pleura, and there is a tendency at this stage for loculations to form within the pleural fluid, which may limit effective drainage of the effusion with a chest tube. The pleural fluid pH and glucose level will be lower, and the LDH level will rise, often dramatically.
The final organization stage is characterized by aggressive fibroblast growth over the pleural surfaces to form an inelastic membrane called the “pleural peel.” This is often extensive and reduces lung functionality considerably. The pleural fluid is often thick, consisting of pus and cellular debris.
Primary empyema instead arises by direct translocation from the oropharynx or by hematogenous spread. In this circumstance, bacteria invade the pleural space as the initial insult leading directly to the fibropurulent stage.
Clinical Presentation
Although there have been some extremely unusual cases, the classic presentation of prolonged pleural infection is difficult to separate from that of pneumonia, whereby patients suffer with dyspnea, cough, fever, malaise, and perhaps pleuritic chest pain. In fact, a significant number of patients with pneumonia will go on to develop a parapneumonic effusion without a change in symptoms to offer clues to its existence. Furthermore, there is no symptomatic discriminator between patients with effusions determined to be uncomplicated or complicated.
An uncomplicated parapneumonic effusion is defined here as an effusion with a glucose level above 40 mg/dL, pH above 7.2, and negative Gram stain and culture without loculation on ultrasonography. A complicated parapneumonic effusion is either loculated on ultrasonography or has a glucose level below 40 mg/dL or a pH below 7.2. A high index of suspicion should be maintained for those patients who fail to improve within a few days of initiating antibiotic therapy or who exhibit persistent fever or signs of sepsis, with further investigations following rapidly. With long-lasting infection, the course of patients with pleural infection can mimic the course of those with malignant processes, often with significant weight loss, sweats, and loss of appetite.
Pleural Fluid Sampling
Abnormal signs, symptoms, or blood test results in the context of a suggestive radiograph should lead to confirmation of the presence of an effusion and early sampling of the fluid. However, in a small retrospective series, Skouras and colleagues suggested that parapneumonic effusions less than 2 cm in thickness on chest CT scan can be treated with antibiotics without sampling because they are unlikely to become complicated or require intervention. Such patients would still require close monitoring and appropriate antibiotic therapy.
Biochemistry
In cases where overt pus is revealed on initial aspiration, no further biochemical analysis is necessary, and chest tube placement is required. Microbiologic analysis is, however, still important in these cases.
Pleural fluid pH may be the best discriminator for a complicated pleural process during initial investigations, with many studies demonstrating better patient outcomes when drainage is instituted based on the early biochemical changes related to infection. In the meta-analysis by Heffner and coworkers, the receiver operating characteristic for the diagnostic accuracy of pleural fluid pH showed that, if the pH was less than 7.2, a chest tube was likely to be necessary to resolve the pleural infection. A pH analysis (or glucose analysis) should therefore be performed in all cases where the diagnosis of pleural infection is being considered. The pH needs to be measured using a blood gas machine because pH strips and pH meters are not accurate enough to be of value. Some institutions rely on the glucose concentration instead, finding it as useful a test and less prone to error.
Current guidelines therefore regard a pH of 7.2 as diagnostic of complexity, making this the definitive “cutoff” below which drainage should take place. It should be noted that fluid pH values may be appreciably altered by minor variations in sampling and processing techniques, which can therefore have significant effects on management strategy. In a study of pleural fluid pH, Rahman and associates reproduced several common scenarios that could result during testing. Even small amounts of residual heparin or local anesthetic in a sample syringe could dramatically lower pH; residual air in the syringe could increase pH (if 1 mL of air was in the syringe with 2 mL of fluid, the pH rose by an average of 0.08). These errors represented a clinically significant change in over two thirds of patients.
As mentioned earlier, when pH measurements are not easily obtainable, a glucose level can be useful. Samples should be collected in a blood glucose tube and sent to the laboratory, with a value of less than 2.2 mmol/L (40 mg/dL) indicating the need for chest tube drainage. It should be noted that pH and glucose level can also be low in malignant pleural effusions, rheumatoid pleural effusions, and pleural effusions secondary to esophageal rupture; therefore their value for guiding chest tube drainage should be restricted to the setting of parapneumonic effusions.
LDH measurements tend to be high in all cases of complicated parapneumonic effusions and empyema. LDH tends to rise rapidly as the pleural infection progresses through the fibropurulent and organizing stages of the disease. When a patient is being treated with antibiotics alone, a rising LDH on repeated thoracentesis may indicate that an effusion is not responding and chest tube drainage should be considered.
Microbiology
The bacteriologic features of culture-positive pleural fluid have changed since the introduction of antibiotics. In the preantibiotic era, the commonest pathogens were Streptococcus pneumoniae or Streptococcus haemolyticus . After this period, between 1955 and 1965, Staphylococcus aureus was the commonest causative organism, with the major shift in the frequency and type of causative organisms attributable to the introduction of antibiotics. There also appear to be global and regional differences in the frequencies and range of the causative organisms.
In a study of patients with pleural infection, standard culture methods were combined with nucleic acid amplification to discern causative organisms, attaining a 74% overall identification rate. Cloning techniques were also applied to a small number of cases (3%), limited by cost. Nucleic acid amplification identified an organism in 38% of the culture-negative samples, with the same organism found by both culture and nucleic acid amplification (or cloning) in 35% of cases.
In this cohort of mostly community-acquired pleural infections (85%), the Streptococcus anginosus group (formerly Streptococcus milleri group) was the predominant species of bacteria. These and other gram-positive aerobes were implicated in 65% of cases, confirming the inherent differences in etiology of empyema compared to pneumonia. Other organisms included staphylococci (11%), gram-negative aerobes such as Escherichia coli (9%), and anaerobes (20%). Polymicrobial samples were identified in 20% of cases, but this may well underestimate the true incidence, as suggested by the results using cloning in the study and the fact that anaerobes, which have been identified in up to three quarters of cases of community-acquired pleural infection in other series, may have been underrepresented in this series.
Hospital-acquired pleural infections made up only 15% of this cohort overall but were quite different from the community-acquired pleural infections. Most (58%) cases were attributed to gram-negative organisms or staphylococci, with over 70% of the latter due to methicillin-resistant S. aureus . A similar gram-negative predominance has been found among patients in the intensive care setting. The bacteriology of hospital-acquired pleural infection is therefore very different from that of community-acquired pleural infections and requires different empirical antibiotics at presentation. Table 80-1 summarizes the frequency of over 2000 culture-positive cases reported in the English literature between 1996 and 2012.
| Organism—Aerobes (Gram-Positive) | Percent | Organism—Aerobes (Gram-Negative) | Percent | Anaerobes | Percent |
|---|---|---|---|---|---|
| Streptococcus pneumoniae | 32 | Escherichia coli | 3 | Fusobacterium | 2 |
| Streptococcus milleri group | 10 | Klebsiella spp | 3 | Bacteroides | 2 |
| Other Streptococcus spp. | 10 | Haemophilus influenzae | 1 | Peptostreptococcus | 6 |
| Staphylococcus aureus | 10 | Other coliform spp. | 2 | Prevotella | 1 |
| Streptococcus pyogenes | 2 | Proteus | 1 | Mixed anaerobes | 1 |
| Methicillin-resistant S. aureus | 2 | Enterobacter spp. | 2 | Other | 6 |
| Enterococcus spp. | 1 | Pseudomonas aeruginosa | 3 | ||
| Total | 67 | 15 | 18 |
The use of blood culture bottles for culturing pleural fluid can increase yield; Menzies and colleagues provided the first comparative, prospective evidence of increased microbiologic diagnostic yield in pleural infection using the BACTEC blood culture bottle system (Becton, Dickinson U.K.). Addition of blood culture bottle–inoculated pleural fluid at the bedside to standard pleural fluid culture increased microbiologic diagnostic yield by 21%, and, in a small proportion of cases (4%) where the standard culture was positive, use of the blood culture bottles suggested the presence of additional organisms that would alter antibiotic management. These findings are further supported by the lack of false-positive culture results from noninfected control samples. The standard cultures were positive in 29% of cases in which the pleural fluid cultured in blood culture bottle cultures were negative, suggesting potential organism preference for certain growth media. This study demonstrates a significant increase in diagnostic yield using a widely available and relatively inexpensive technique, suggesting that inoculation of pleural fluid into blood culture bottles should be added to standard pleural fluid culture routinely.
Antibiotic Selection and Duration
The initial empirical antibiotic selection at presentation should be based on whether the pleural infection is community acquired or hospital acquired. Hospital-acquired pleural infection has a different range of causative organisms and a considerably higher mortality ( Fig. 80-2 ).
Because anaerobic bacteria are commonly seen in this setting and often fail to grow in culture, they should be covered empirically in all cases of presumed pleural infection. Liaison with local microbiologists is invaluable in deciding the best combination of antibiotics to use, because resistance patterns vary significantly from one geographic region to the next.
There are no robust studies looking at the optimal duration of antibiotics for pleural infection. In our experience, if a patient is ill enough to require hospitalization, then intravenous antibiotics are usually indicated. When converted to an oral equivalent, perhaps before discharge, a further 1 to 3 weeks of antibiotics is commonly needed. The length of the antibiotic course is governed by the inflammatory markers ( C-reactive protein [CRP]) and any ongoing, often mild fever, rather than by the radiologic appearance, which often lags behind clinical improvement.
The potential role of intrapleural antibiotics is not often considered and randomized, controlled trial data are lacking. Because most intravenous antibiotics penetrate into the pleural space in suitable concentrations, there is no current role for intrapleural antibiotic use in this setting. However, it should be noted that aminoglycosides do not penetrate the pleural space well and should be avoided.
Nutrition
Patients with pleural infection, particularly those with empyema who have had a delayed presentation, suffer the protracted catabolic consequences of chronic infection. A low albumin level has been shown to be a marker of poor outcome in one large published series. Addressing the patient’s nutritional status at presentation is often overlooked and should be an early priority alongside tube drainage and the prescribing of suitable antibiotics. Early nutritional assessment should be seen as mandatory.
Early Risk Stratification
A reliable and sensitive clinical prediction model of poor outcome in pleural infection would enable clinicians to triage patients according to risk and to select the more aggressive and expensive therapies for patients who may otherwise have the poorest outcomes. To date, there are no robust validated methods for identifying high-risk patients at presentation with pleural infection. Selection for surgery has previously been based on duration of symptoms, pleural fluid purulence, size of infected pleural fluid collection, and the degree of parietal pleural thickening on imaging. In a cohort study of 85 sequential patients, clinical care was based on structured treatment guidelines to assess whether the generally accepted baseline predictors reliably identified patients at high risk. Only pleural fluid purulence had predictive power for a poor outcome, and this was insufficiently sensitive and specific to be of clinical value. In a second study, this finding was confirmed and predictors of residual pleural thickening were identified, although thickening was uncommon and not associated with clinical disability.
Although there are likely to be complex interactions between genetic and environmental factors that contribute to the development of pleural infection, these have not yet been determined. There are, however, certain patient risk factors, particularly chronic alcohol excess and intravenous drug use, which likely increase the risk because of aspiration of gastric contents. In addition to these, Chalmers and colleagues described four other independent risk factors that seem to predict the development of pleural infection: serum albumin level below 30 g/L, serum CRP greater than 100 mg/L, platelet count above 400 × 10 9 /L, and serum sodium level below 130 mmol/L. This study noted than none of the routinely employed pneumonia or sepsis scores were adequate in determining this outcome and suggested a score based on these six factors, although this still requires validation. Interestingly, patients with chronic obstructive pulmonary disease were found to be at lower risk for developing pleural infection, perhaps due to a background level of generalized inflammation causing an attenuated response to a pleural bacterial challenge.
For patients with confirmed pleural infection, those at greatest risk for poor outcome may be identifiable. For the patients recruited to the U.K. MIST1 trial, an outcome score was developed and subsequently validated using a second cohort from the MIST2 trial. Of the 32 baseline characteristics analyzed, 5 presenting factors (age, serum urea level, serum albumin level, fluid purulence, and likely origin of infection) could predict the eventual outcome, with patients divided into low-, medium-, or high-risk groups. Patients in the lowest risk group were found to have a mortality of less than 5% at 3 months, whereas those in the highest were found to have a mortality approaching 50% over the same time period. The main potential advantage of this stratification system is in allowing physicians to institute fibrinolytics or surgery earlier in the clinical presentation when they are perhaps more likely to be successful.
Imaging Techniques for Pleural Infection
Radiologic tests are vital in the initial diagnosis and management of pleural infection. Chest radiography, CT, and ultrasonography are all useful tools in the management of patients with pleural infection.
Radiology
The presence of fever, pleuritic pain, and pleural effusion should always alert the physician to the possibility of pleural infection. Pleural fluid loculation can result in a D-shaped subpleural opacity ( Fig. 80-3 and eFig. 80-1A ) on chest radiography, which could be misinterpreted as a lung mass if one is not aware of this common appearance. In ventilated patients in the supine position, free fluid may track posteriorly and cause haziness of the hemithorax on the chest radiograph.
Thoracic ultrasonography (see eFig. 80-1 ) is advisable in all cases of suspected pleural infection because it can provide important information and it allows accurate chest tube placement into the most appropriate part of the pleural collection. Use of ultrasonography has also been shown to reduce iatrogenic injury. Ultrasonography is also more accurate than CT imaging in detecting loculations and septations and can detect the presence of small quantities of pleural fluid not visible on chest radiography.
Chest CT (see eFig. 80-1H–K ), particularly contrast-enhanced chest CT ( eFig. 80-2 ), is not necessary in all patients with pleural infection and should be reserved for those with persistent collections despite attempted chest tube drainage, those who may have a suspected proximal obstructing lesion, and those in whom surgery is being considered. It will provide detailed information about fluid loculation ( 
), identify chest tube position, and differentiate between empyema and lung abscesses (see eFig. 33-7B–E ) when there is diagnostic uncertainty ( Fig. 80-4 and Table 80-2 ).
| Empyema | Lung Abscess |
|---|---|
| Lenticular shape | Rounded |
| Surrounding lung often compressed | Boundary between lung and fluid indistinct |
| Margins of collection create obtuse angles with chest wall | Makes contact with chest wall at acute angle |
| Thick smooth wall | Thick irregular wall |
| No vessels close by | Vessels seen passing through or near collection |
Magnetic resonance imaging is usually reserved for those patients who are unable to undergo CT imaging or for persons at high risk from irradiation. It is very good at visualizing septations and loculations within the pleural fluid ( Fig. 80-5 ).
Bronchoscopy
Practitioners should be aware that a proximal obstructing lesion may be the cause of an empyema. Although uncommon (less than 4% in one large series), it should be considered in patients with a suggestive plain chest radiograph, especially one in which the mediastinum is shifted toward the side of the effusion, or in those failing to respond to simple first-line management. A bronchoscopy and chest CT would be the investigations of choice if a bronchial obstruction is suspected. There is no role for routine bronchoscopy in all patients with pleural infection.
When an empyema is discovered distal to an obstructed bronchus, the underlying cause is usually malignancy. Once histologic confirmation has been obtained, relief of the obstruction with radiation therapy, laser resection, or stent insertion might allow the empyema to be treated effectively. Where this is not possible, prolonged oral antibiotics may be useful to prevent ongoing sepsis.
Tube Size
Over the past decade there has been a shift away from large-bore catheters to small-bore chest tubes inserted by the Seldinger technique. This change in practice is not so much the result of large, well-conducted, randomized controlled trials but more to do with the relative ease of the technique and the benefits of the smaller drains, which cause the patient less pain during insertion and while in situ.
For cases of pleural infection, we recommend initial placement of a small-bore (12-French) chest tube under ultrasonographic guidance and then management with regular saline flushes to ensure that it does not become occluded. Guidelines recommend the instillation of 30 mL saline every 6 hours via a three-way stopcock. If this fails to drain the effusion effectively, a large-bore drain may occasionally be necessary. It is more common for tubes to be replaced not because of obstruction but because they have become dislodged or fall out as a result of failure to secure the chest tube correctly.
Although there are no randomized data comparing large-bore with small-bore chest tubes for pleural infection, one study reviewed the 405 patients who participated in the MIST trial. As part of the study, data were collected about the size of chest tube used. No difference between small- and large-bore tubes was found in the frequency with which patients either died or required thoracic surgery. However, those with larger chest tubes reported significantly more pain both at the time of insertion and while the drain remained in situ.
In another study of 103 patients with pleural infection, the small-bore chest tubes provided definitive treatment in 78% of cases, similar to the success rates of two other series in which large-bore chest tubes were used. It should be noted that, in this study, radiologists guided tube placement, which almost certainly helped the result. With this in mind, we recommend that chest tubes should be placed under ultrasonographic guidance in the largest locule in a dependent part of the pleural effusion.
Fibrinolytic Therapy
The use of fibrinolytic agents to disrupt fibrinous pleural septations was first described by Tillet and Sherry in 1949, who used partially purified streptococcal fibrinolysin that contained both streptokinase and streptodornase (a DNase) to drain infected postoperative hemothoraces. It was associated with immunologic side effects and did not become routine practice.
Streptokinase is a proteolytic enzyme derived from a bacterial protein of group C β-hemolytic streptococci. It forms a complex with plasminogen that then converts additional circulating plasminogen to plasmin. Plasmin lyses fresh fibrin clot and digests prothrombin and fibrinogen. Because it is derived from a bacterial source, it is antigenic, unlike urokinase.
In a series of 24 patients, Davies and coworkers began to allay fears regarding intrapleural thrombolysis, which was found to be safe and also led to improvements in clinical outcomes. Other studies tended to focus more on the use of urokinase in loculated effusions, with benefits being demonstrated in terms of a reduction in treatment failure (as judged by surgical referral or death), reduced length of hospital stay, and better surgical outcomes. However, these studies were often small trials or case series, which limited their ability to be generalized.
The 2005 MIST1 trial recruited 454 pleural infection patients from the United Kingdom to receive either intrapleural streptokinase or placebo. Entry criteria reflected real-world practice with a strong reliance on local physician diagnosis, antibiotic choice, chest tube use, and surgical referral. The trial was unable to show any significant benefit from the use of streptokinase in either surgical referral or death. There was also no improvement in the length of stay or in any signal in any subgroup analysis. However, a Cochrane review found that intrapleural fibrinolytics conferred benefit both in reduced treatment failure and reduced need for surgical intervention in pleural infection, but not reduced mortality. The choice of streptokinase as the primary lytic may have contributed to these results, because its mechanism of action relies upon using a proportion of the intrapleural plasminogen to form an active complex before the rest can be converted to plasmin. Nonetheless, the 2010 British Thoracic Society guidelines advised that intrapleural fibrinolytics should not be used routinely, but might be considered in select cases.
Following this, in the blinded, 2-by-2 factorial MIST2 trial, 210 patients with pleural infection were randomly assigned to receive one of four study treatments for 3 days: intrapleural tissue plasminogen activator (t-PA) and DNase, t-PA and placebo, DNase and placebo, or double placebo. The primary outcome was the change in pleural opacity, measured as the percentage of the hemithorax occupied by effusion on chest radiography on day 7 as compared with day 1. The mean (±SD) change in pleural opacity was significantly greater in the t-PA–DNase group than in the double placebo group (−29.5% ± 23.3% versus −17.2% ± 19.6%; difference, −7.9%); the change observed with t-PA alone and with DNase alone (−17.2% ± 24.3% and −14.7% ± 16.4%, respectively) was not significantly different from that observed with placebo. The frequency of surgical referral at 3 months was lower in the t-PA–DNase group than in the double placebo group (2 of 48 patients [4%] versus 8 of 51 patients [16%]; the odds ratio for surgical referral, 0.17, and was greater in the DNase group (18 of 46 patients [39%]) than in the placebo group (odds ratio, 3.56). Combined t-PA–DNase therapy was associated with a reduction in the hospital stay as compared with double placebo (difference, −6.7 days); the hospital stay with either agent alone was not significantly different from that with placebo.
The authors concluded that intrapleural t-PA–DNase therapy improved fluid drainage in patients with pleural infection and reduced the frequency of surgical referral and the duration of the hospital stay. Treatment with DNase alone or t-PA alone was ineffective; indeed, DNase monotherapy was associated with an increase in surgical referrals by a factor of 3. Reduction in the infected pleural fluid collection was roughly doubled with the use of the combination therapy (with clearing of approximately 30% of the ipsilateral hemithorax volume and about a 60% reduction in the baseline pleural collection). This treatment was not associated with an excess of adverse events. In a report from eight centers, the use of tPA/DNase also showed benefit, although with some pain and bleeding, in those who did not respond to antibiotics and thoracostomy drainage. In summary, this approach looks promising, although a larger trial is needed to confirm safety and guide physicians on which patient groups are likely to gain the most benefit.
Monitoring Response to Medical Management
Identifying patients who are not responding to medical management can be a challenging task. Improvements in the imaging appearance often lag behind clinical improvement and therefore, in our opinion, are not a good indicator of the need for further intervention. The best markers of response to medical management include a falling CRP (halving in value and ideally falling to below 100 mg/L), the settling of a spiking temperature, and clinical signs associated with resolution of sepsis. If these markers are all improving, then further interventions with additional chest tubes or surgery rarely becomes necessary.
When patients look as if they have avoided surgery by the narrowest of margins, prolonged oral antibiotics (for 4 to 6 weeks) and close follow-up as outpatients with monitoring of the chest radiograph and CRP is advised.
Surgical Options
Patients are usually referred for surgical intervention after initial medical treatment failed or if they presented late with highly organized empyemas that demonstrate significant pleural thickening and loculation. Practice varies, with some centers having an extremely low threshold for early surgery. The point at which medical management is deemed to have “failed” is necessarily ill defined, but one important indicator would be signs of ongoing sepsis despite attempted chest tube drainage and adequate antibiotic therapy. Another important consideration is the risk for long-term respiratory embarrassment without the removal of the fibrin and loculated fluid.
Modern surgical options are varied and may be tailored to the individual. VATS typically requires general anesthetic and single-lung ventilation but can be performed under local anesthetic in patients deemed too high risk for a general anesthetic ( Fig. 80-6 and 
). Although originally used for thorough pleural débridement, VATS can now be used to perform decortication in particularly advanced or chronic empyema, although this latter situation may reduce the chance of successful outcome. Despite this, overall success rates for VATS, measured by resolution of sepsis and clinical stability, exceed 85%.
