Interventional radiology treatment of empyema and lung abscesses

Introduction 

Pneumonia in children is quite common and usually responds either to oral or intravenous antibiotics. However, in a subset of patients, the pneumonia can be complicated by pleural effusion, empyema or necrotization progressing to pneumatoceles or abscesses. These children often require additional interventional therapies to resolve their primary illness, with a secondary goal of preserving lung function.1 These complications are most likely related to the virulence of the pathogen, with Streptococcus pneumoniae being the most common bacterium.2, 3, 4, 5 The incidence of complicated pneumonia due to pneumoccocal disease is increasing.2, 3

It is well established that appropriate treatment of empyema will help control sepsis, restore pulmonary function and prevent lung entrapment from the fibrous peel.2, 6, 7 Traditionally, pleural collections were treated with large bore catheters inserted by a surgeon. If this failed to completely drain the pleural space, surgical intervention with a decortication was performed. Later developments included instilling a fibrinolytic into the chest tube in an attempt to break up loculations and video-assisted thoracoscopic surgery (VATS).1, 8, 9, 10, 11 Interventional radiologists originally had a limited role in the management of these patients, with image guided procedures indicated if residual loculated collections remained after surgical therapy. However, after successful treatment of these post-surgical patients, image guided interventional techniques were adapted for primary treatment of patients with pleural collections, as they were better tolerated than surgery, avoided general anaesthesia, had a high clinical success and were without significant complications.

Abscess and pneumatocele complications from pneumonia occur less often than does pleural infection, and while most commonly due to pneumococcal disease, polymicrobial infection or other pathogens have a higher incidence than with empyema. In addition, congenital malformations such as sequestration or congenital cystic adenomatoid malformation may present as pulmonary abscesses.12 In the majority of patients, the abscess spontaneously drains via the airways and prolonged antibiotic therapy is sufficient. Spontaneous drainage is less common is patients younger than 7 years old, with up to 21% failing medical treatment.13 Similar to pleural disease, this problem was traditionally managed surgically and usually required a partial or complete lobectomy, with the potential of reducing later pulmonary function. Interventional radiology is the preferred treatment modality for abscesses in most other areas of the body. These same techniques have been adapted to treating intrapulmonary collections, with a high degree of success.

The purpose of the article is to describe image guided techniques for treating complicated pneumonias and to evaluate the role of these techniques in the care of paediatric patients.

Interventional radiology techniques 

Empyema 

Imaging 

Chest radiographs are usually the first imaging to be done in a patient with pneumonia (Fig. 1a). Advanced imaging may not be needed, but is often helpful. The amount of pleural fluid is often not apparent on the chest film as an opaque hemithorax may have little fluid or the entire lung may be collapsed by a large empyema.14

View Large Image Download to PowerPoint

Figure 1. This patient presented with a right upper lobe pneumonia (*) and a pleural effusion (arrow) seen in (a). The chest CT (b) shows the effusion (*) which appears to be free flowing as it is dependent. An ultrasound in (c) shows multiple septations in the pleural fluid (arrows). (d) shows the radiograph after image guided insertion of a small bore chest tube and fibrinolytic therapy. The empyema is nearly resolved, with persistent pneumonia (*) causing persistent fevers. After removal of the chest tube and completion of antibiotics, (e) shows minimal residual pleural thickening (arrow).

A CT scan will show all areas of pleural disease and can direct tube placement; however, CT is poor at distinguishing the character of the fluid (transudate versus exudate) or the presence of loculations2, 14, 15, 16 (Fig. 1b).

Ultrasound imaging can verify or exclude the presence of fluid. It also grades the fluid, depending on the identification of loculations.2 Stage I is anechoic fluid, Stage II has loculations and Stage III is identified to be a solid peel.17 However, ultrasound is a directed study and distant sites of disease may be missed14 (Fig. 1c).

Procedure 

Almost all patients will require some form of sedation or anaesthesia. Standard sedation protocols are followed. We typically use a combination of pentobarbital and fentanyl intravenously in children younger than 12 years old, with midazolam substituting for pentobarbital in older patients. General anaesthesia or other medications may be used as well, when available.

While some advocate serial thoracentesis for thin effusions,7 we only perform thoracentesis for diagnosis of the pathogen. If a patient has fever or oxygen requirement, we always place a tube. This eliminates the need for multiple sedations and it is easier to remove the tube than to perform the procedure again. The patient is placed supine or with the affected side slightly elevated. The arm is either raised over the head or placed across the anterior chest. Pre-procedural ultrasound scanning is performed to determine the best site for access, the most appropriate transducer and the character of the fluid. Typically, we use an 8MHz curvilinear transducer with a small footprint as it easily fits into the small imaging window between the ribs. A standard sterile prep and drape are performed.

The goal is to place the tube in a posterior location as the fluid is dependent. The mid-axillary line is usually chosen as it allows for posterior tube placement without the patient having to lie on the tube. For isolated collections, the most appropriate access to the empyema is chosen based on imaging findings. Local anaesthesia is provided, and with direct ultrasonographic guidance, a sheathed needle is directed posteriorly and over the rib to avoid the intercostal vessels. A small sample of fluid is taken and sent to the laboratory.

The remainder of the procedure is performed with fluoroscopic guidance. A guidewire with a floppy tip is advanced and directed posteriorly. Over the wire, the tract is dilated and a pigtail catheter placed. We generally use 12-French catheters, although the size may be decreased in smaller children or increased if the ultrasound characteristics suggest thick fluid. Alternatively, other catheters may be used. The wire or catheter can be manipulated through the pleural space to try and break up adhesions; however, this may lead to wire or catheter kinking and does not usually result is a large amount of additional drainage compared to the use of fibrinolytics. The catheter is secured to the patient and is attached to a standard chest drain with 20cmH2O suction.

Fibrinolytics 

Originally, we reserved the use of fibrinolytics for when ultrasound showed multiple loculations, or when tube output dropped and imaging showed significant residual fluid; however, we believe this lead to some failures. Now, unless the fluid appears serous and is purely anechoic without loculations (rare), every patient is started on a fibrinolytic. We currently use tissue plasminogen activator (tPA) 2mg for all patients, mixed in 20ml of saline. The tPA is injected into the tube, the tube is clamped for 1h and then the tube is returned to suction. This is repeated every 12h until fluid output decreases to <25ml in a 12-h period.

Follow-up 

Tube output, fever and oxygen requirement are followed, and a chest radiograph is obtained daily (Fig. 1d). At 48–72h, if the patient has not improved, and residual fluid is suspected from tube output less than expected (from initial imaging) or there is a worsening chest radiograph, repeat imaging is performed. If significant pleural disease is present, then tube drainage with fibrinolytics is considered a failure and the patient is referred for surgical intervention. Otherwise, tPA is continued and the tube is removed when output falls and one of the two following criteria is met:

1.The patient is afebrile and off supplemental oxygen, or

2.Imaging has demonstrated no significant pleural disease.

We do not routinely obtain follow-up CT or ultrasound imaging if these criteria are met (Fig. 1e).

Results 

The results of image guided small bore catheter drainage with adjuvant fibrinolytics have been reported in a number of series and these are summarized in Table 1.18, 19, 20, 21, 22, 23, 24, 25

Table 1. Results of image guided small bore catheter drainage with adjuvant fibrinolytics for empyema18, 19, 20, 21, 22, 23, 24, 25

	Study	Success rate	LOS (days)
	Gates et al.18	24/29 (89%)	9.3
	Wells & Havens19	70/71 (99%)	11.2
	Yao et al.20	18/20 (90%)	16.6
	Ekingen et al.21	26/32 (81%)	19.9
	Sonnappa et al.22	25/30 (83%)	6.0
	Cochran et al.23	14/19 (74%)	13.0
	Khalil et al.24	14/17 (82%)	11.0
	Feola et al.25	54/58 (93%)	9.1

LOS, length of stay.

Complications 

Theoretical complications from image guided drainage include bleeding, vascular injury, bronchopleural fistula (BPF), anaesthesia or sedation complications, and death. However, these are rare. There is only one published report of a death from haemorrhage.2 There have been some reports of BPF development after drainage and fibrinolytics. However, distinguishing an iatrogenic BPF from a pre-existing or developing air leak from lung necrotization is difficult.

Abscesses 

Indications for drainage include persistent symptoms despite appropriate medical treatment, and failure of resolution on imaging after a course of antibiotics. In addition, there is an operative risk of spillage from the abscess into the normal lung via the airways when the patient in placed in the decubitis position for thoracotomy. Pre-operative drainage is thus often requested.

Imaging 

Intrapulmonary infected collections require cross-sectional imaging. CT can identify the abscess, determine the number of collections, identify its position relative to the visceral pleura, vasculature and airways, determine if a BPF is present and evaluate the pleural space for concomitant empyema (Fig. 2a). Ultrasound is of limited use in pre-procedural planning as interposed air makes identification of the abscess difficult.

View Large Image Download to PowerPoint

Figure 2. (a) Shows a cavitary lesion in the left lower lobe (*) with an air fluid level representing an abscess. After insertion of a chest drain with image guidance (b), a follow up CT shows a markedly smaller abscess (arrow) with the drain in the middle.

Procedure 

The technique for image guided abscess drainage is similar to that for empyema, with some key differences. The site of access and path for tube placement are more critical. Tube placement across normal lung or an interlobar fissures is avoided if possible. While ultrasound guidance is used whenever possible, CT or fluoroscopy may need to be used instead if the sonographic window is poor (Fig. 2b).

Fibrinolytics 

Fibrinolytics are not usually indicated for pulmonary abscesses although, if the collection is loculated, they may be used in an attempt to get complete drainage via a single catheter.

Follow-up 

Follow-up for abscesses is also similar to that for empyema. In our experience, the duration of drainage is about the same, unless a BPF occurs. With a BPF, longer drainage with a higher suction pressure may be required. The catheter is removed when drainage slows, symptoms improve and the abscess decreases in size based on a chest radiograph. As with empyema, we discourage additional CT imaging in a patient with clinical improvement for radiation protection reasons.

Results 

The results of image guided small bore catheter drainage with adjuvant fibrinolytics have been reported in a number of series, as summarized in Table 2.12, 26, 27, 28, 29, 30

Table 2. Results of image guided small bore catheter drainage with adjuvant fibrinolytics for pulmonary abscesses12, 26, 27, 28, 29, 30

	Study	Success rate
	Kosloske et al.26	3/4 (75%)
	Zuhdi et al.27	6/6 (100%)
	Hoffer et al.28	4/5 (90%)
	Patradoon-Ho & Fitzgerald12	19/19 (100%)
	Imamoglu et al.29	5/7 (71%)
	Ball et al.30	8/8 (100%)

Complications 

Potential complications are the same for empyema drainage, with the incidence of BPF ranging from 0 to 75%.26, 27, 28, 29, 30 However, many abscesses already communicate with the airway and BPFs should be expected. The overall complication rate is low, with no reported deaths.

Discussion 

Multiple reports have cited the incidence of pleural effusion with pneumonia to be 36–57%, with 15–20% progressing to empyema, defined as pus within the pleural space.2, 4, 6, 7, 15, 21, 31 Empyema is classically divided into three separate stages:5, 6, 16, 31, 32, 33, 34, 35

1.Early-exudative: lasts from 24 to 48h; fluid forms due to pleural inflammation.

2.Fibrinopurulent: lasts from 7 to 10 d; characterized by thickened exudates, fibrin deposition, white cell infiltration and formation of loculations.

3.Organized: starts 2–4 weeks from effusion onset. There is fibroblast formation and creation of a thick peel.

Grading of pleural disease based on laboratory criteria has also been described, with pH, glucose and lactose dehydrogenase (LDH) levels, Gram stain and cultures all being performed.2, 7, 21 However, laboratory evaluation does not define how the patient should be treated, with the exception of antibiotic coverage for the appropriate pathogen.12

Several factors have prognostic value in determining if interventional therapy will be required. These include duration of symptoms, co-existing medical conditions and ultrasonographic stage.5, 1, 10, 36 Early intervention can improve symptoms, decrease the length of morbidity and reduce complications. Serial thoracentesis has been proposed; however, this requires multiple invasive procedures in a child who might require sedation.2, 31 Alternatively, early tube drainage while the fluid is simple can be effective.

VATS has been recommended as a first-line approach in patients with empyema.1, 5, 8, 9, 10, 11 This has been shown to be successful with high curative rates and a low complication rate. The primary measured outcome is length of stay (LOS) with improvement compared to traditional therapies. However, VATS is still an operative procedure requiring general anaesthesia. While this therapy may be necessary in some patients, many of these children could be successfully treated with a less aggressive approach. Therefore, alternative treatments have been explored.

Image guided small bore tube placement with adjuvant fibrinolytic therapy also has high success rates and low complications rates.18, 19, 20, 21, 22, 23, 24, 25 At institutions with interventional radiology facilities to perform image guided small tube placement and fibrinolytic therapy, success rates are 74–99%.18, 19, 20, 21, 22, 23, 24, 25 The LOS in multiple studies is similar to VATS.

Imaging of patients with pneumonia can help determine if complications such as empyema, abscess or BPF are present, or if interventional therapy is needed, and can also direct therapy. When empyema is present, CT is the most efficient way to identify every area of involvement and can guide where a catheter needs to be placed. In our experience, CT is most valuable in evaluating the parenchymal disease. Necrotization is readily identified and the full extent of the pneumonia is seen. This provides prognostic information about the duration of medical therapy and possible need for future interventions such as abscess drainage. In addition, in many cases where patients have remained ill despite interventional radiology empyema drainage, CT has shown that the problem was progressive necrotizing pneumonia instead of failure of pleural infection therapy. While a CT scan prior to empyema treatment gives valuable prognostic information (primarily about the necrotizing pneumonia), it can be deferred due to radiation concerns. However, a CT scan is mandatory before declaring interventional radiology pleural drainage a failure, and sending a patient to the operating room for VATS or thoracotomy.

Ultrasound is valuable in these patients. Ultrasound grading of the empyema is prognostic, with Grades II and III requiring a greater level of intervention.36 While Grade I fluid is thin and easily removed by thoracentesis, tube placement may be in the patient's best interest to avoid multiple procedures with sedation. Due to loculations, the higher grades will not completely drain via simple tube placement; therefore, fibrinolytics are indicated for these patients. In addition, empyema is a dynamic process2 and a Grade I pleural collection may change character over time as the visceral pleura continues to secrete fluid as a response to the pneumonia. In our early experience,18 some failures of tube drainage occurred when Grade I fluid treated without fibrinolytics became loculated over time, or when we stopped fibrinolytic therapy early, with later peel formation. In our practice, for Grade I empyemas, we start fibrinolytic therapy immediately if the tube output is below expected (as predicted by imaging), or within 12h if the chest radiograph is not completely clear of pleural disease. Fibrinolytics are always given if there are loculations on ultrasound (Grades II and III). Grade III empyemas have a greater failure rate with interventional techniques, although others find no difference in these patients compared to surgery. Therefore, if ultrasound shows what appears to be a fibrous peel, we consult surgery. Some of these patients will respond to interventional techniques, and they can be attempted with the knowledge that surgery will not be delayed if the patient does not improve expediently. The limit of ultrasound is that it is a focused study; not all areas of the pleura can be evaluated. Therefore, the risk is that some areas of loculated empyema could be missed if a CT scan is not performed. This may not make a difference, as the fibrinolytics break up these loculations; however, this reiterates why a CT scan is important in the follow-up of a patient who remains ill despite therapy.

Fibrinolytics significantly increase the amount of pleural fluid amenable to drain out of small tubes. The primary benefit of this is the access gained to collections that are walled off by septations. Contrary to some perceptions, the fluid in loculated empyemas usually remains thin. It can easily be drained by a tube if not isolated by the loculations. The type of fibrinolytic used has not been shown to make a significant difference. Urokinase was most commonly used, but conversion to streptokinase or tPA was required in the USA when the Food and Drug Administration (FDA) removed urokinase from the market. Most institutions have stopped using streptokinase due to concerns over possible immunological reactions. There are both weight-based (0.1mg/kg, up to 3mg) and set doses (2mg) for tPA; however, the efficacy of each appears similar, and no pleural bleeding requiring transfusion has been identified.18, 19, 25 The dose of urokinase is typically 40000–100000U. All of these medications are mixed with normal saline to an appropriate volume, injected into the tube and left in the pleural space for approximately 1h before resuming drainage.

While image guided small bore tube drainage with fibrinolytics has a similar success rate to VATS, the results are not absolute. Some patients will fail this less aggressive treatment. Even with VATS, some patients are converted to thoracotomy based on the operative findings. Since the LOS for the patient has been proposed as an appropriate outcome indicator for results, treatment algorithms should be tailored to the goal of providing the least aggressive therapy to improve symptoms as quickly as possible. Several studies comparing VATS to tube drainage (with or without fibrinolytics) have concluded that VATS is better due to a reduced LOS. However, a problem with these papers is that in some patients the tube drainage was continued for long periods of time, despite having not succeeded in reducing symptoms in a timely manner. In our practice, if the patient is not improving within 48–72h, we consult with surgery to determine if VATS or a thoracotomy is indicated. This decision is based on clinical parameters such as fever and oxygen requirement. Imaging is also obtained to determine if there is significant residual pleural disease, as surgery is not indicated if it is absent. In most patients with continued symptoms, significant pulmonary disease remains, often with necrotization. This approach of attempting to identify pleural peel (Grade III) with ultrasound, early fibrinolytic treatment and reassessment at 48–72h has improved our success rate to >95% (unpublished). In addition, tube drainage and surgery are not mutually exclusive.22, 23 Many patients requiring VATS have had a trial of drainage via a surgically placed tube, and patients receiving VATS are referred for interventional radiology drainage or open thoracotomy if they do not improve.

Infected lung collections can be primary from a necrotizing pneumonia, or secondary in a patient with a predisposing condition such as a congenital lung lesion. The incidence of abscesses, like empyema, is increasing due primarily to the increased virulence of pneumoccocal disease. Historically, these cases were treated with a prolonged course of antibiotics. There are several case series, dating back to the 1980s, of image guided drainage of these abscesses,26, 27, 28, 29, 30 performed when antibiotics failed or to ameliorate the symptoms expediently. The experience in adults is greater, but similar techniques and success rates have been found in children. As an alternative to thoracotomy and lobectomy, interventional radiology can treat these patients if the clinical symptoms warrant. In addition, pre-operative drainage of infected congenital lesions decreases the risks of pulmonary failure and overwhelming sepsis which can occur due to bronchial spillage during surgery.

The incidence of complications from interventional radiology drainage of pleural disease and intrapulmonary collections is low, with BPFs being the most common. BPFs are also identified after VATS. The development of a BPF is most likely a result of the necrotizing pneumonia causing pleural disruption, as it is rarely seen without concomitant lung necrosis. Theoretically, pleural damage can occur with tube placement, or the fibrinolytic could break up a loculation that is sealing a pre-existing contained air leak. Regardless of the cause of the BPF, treatment is indicated for patient's symptoms, they occur despite the therapy provided, and they can usually be treated with a course of prolonged drainage.

Conclusion 

Interventional radiological techniques can successfully treat empyemas and abscesses complicating pneumonia in children. These therapies are less invasive, have a similar LOS and have similar success and complication rates as compared to operative techniques such as VATS or thoracotomy. Treatment algorithms can be developed to expedite patient care. Less aggressive techniques should be employed first, with rapid referral to surgery if the patient does not improve. With this approach, most patients can be treated with image guided techniques, with the occasional failure being referred to the operating room expediently, and the LOS kept within acceptable parameters.

Practice Points 

Research directions 

Educational Aims 

CME section 

This article has been accredited for CME learning by the European Board for Accreditation in Pneumology (EBAP). You can receive 1 CME credit by successfully answering these questions online.