Empyema

"Pleural disease" is a Level 1 topic from Section 2.1.5 in the first edition of the CICM Syllabus for the Second Part Examination, where the college leaves the question "which disease" sufficiently open for the same syllabus item to embrace empyema, pleural effusion, and theoretically pneumothorax or even mesothelioma. 

Empyema appears in the CICM exam mostly in the form of a differential for a persistent fever or a patch of radiological opacity. It became the topic of an SAQ for the first time in Question 4 from the second paper of 2022, which was half about drainage options, and half about the differences between vancomycin and linezolid for a specific indication.

For that rare exam candidate with enough time to read anything other than college exam papers and UpToDate, Bryant & Salmon (1996) give an excellent overview of the pathophysiology, and for management questions, one could go to either the 2017 AATS guidelines, or, because we are a colony, to the 2022 BTS guidelines for the management of pleural disease. Another good review article that summarises all the relevant evidence is Hu et al (2021). 

Aetiology of empyema

"Empyema" just means "pus in a cavity", etymologically derived from puon (pus) and em (in something). The term is not exclusively restricted to the collections that form in the pleural cavity, and you can get subdural empyemas, gall bladder empyemas, uterine empyemas, bladder empyemas, joint empyemas and literally any other kind of word+empyema combinations where the author of the paper decided to use "empyema" instead of "collection" to describe the presence of pus in some previously empty real or potential space. The distinction between these terms appears to be entirely arbitrary and their patterns of use seem to depend mostly on local culture. Still, as the only college question to ever deal with this was asking about a pleural MRSA collection, from here forward the word "empyema" will mean "an infected pleural effusion".

So: this can form for a variety of reasons, which can be broadly separated into two main groups:

• Fluid collected in the pleural cavity somehow, and then became infected
• An infection formed in the pleural cavity, and attracted a lot of fluid

Either way, the purulent pleural collection  progresses through some predictable stages, the classification of which was created in the 1960s:

• Stage 1. Simple exudate. It is just an exudate with a decent amount of neutrophils in it, in other words pus. "Uncomplicated parapneumonic effusion" is occasionally the term used to refer to this sort of collection, implying that it is somehow simple and carefree.
• Stage 2. Fibrinopurulent stage. It is pus, but heterogeneous in texture, and clearly infected (positive gram stain, etc). The term "complicated parapneumonic effusion" is used to describe this situation, implying that the relationship has transitioned from romantic comedy to drama. A creamy "frank pus" appearance gives the person doing the drainage a right to call this "empyema".
• Stage 3. Organising stage. It is developing into loculations, with potentially rather diverse conditions in each pocket.

To this one could theoretically add "Stage 0: complicated parapneumonic effusion" as this is often the antecedent of Stage 1. These classifications remain relevant since the mid-20th century because each stage has slightly different management options available.

Is this even empyema?

How is it not just a pleural effusion of some benign cause? 20-40% of all inpatients with bacterial pneumonia end up developing effusions, but most of them don't go on to require decortication or chest drainage, so surely some of these patients don't have empyema per se, but just a pleural effusion with some increased protein and cellularity ? Well: features strongly suggestive of "proper" empyema include:

• Loculations
• Thickened pleura
• Enhancing pleura on CT with contrast
• Obviously, positive pleural fluid and gram stain

An uncomplicated parapneumonic effusion (i.e. one which is exudative but not infected) will likely have none of these features, and will usually resolve with antibiotics alone (i.e. the effusion goes away with the pneumonia). About 90% will do this, and the remaining 10% will progress to become an empyema, according to Strange & Sahn, 1999.

When one hears "empyema", one typically lunges for the antibiotics, but in fact it appears that there may be causes that are not ostensibly infectious, and there are mentions in the literature of empyemas which are due to inflammatory nonifectious causes. From Parta (2020), in "Mandell, Douglas, and Bennett's Principles and Practice of Infectious Diseases":

• Rheumatoid arthritis
• Ankylosing spondylitis
• SLE
• Malignancy (eg. from a primary pleural malignancy like mesothelioma, or from necrotic mediastinal nodes)
• Pancreatitis
• Decomposing haemothorax
• Eosinophilic empyemas (Behçet syndrome, Churg-Strauss, etc)
• Meigs syndrome

Infectious causes of empyema

• S.aureus
• S. milleri
• S.pneumoniae
• Enterobacter sp.
• Enterococci
• Klebsiella sp
• Pseudomonas aeruginosa
• Anaerobes in 11-17% of cases

This is from a table in Addala et al (2021). Additionally, Dyrhovden et al (2019) lists a selection of organisms which were found unexpectedly and which are mostly of periodontal origin (eg. Fusobacterium nucleatum and Streptococcus intermedius), presumably spread there haematogenously.

Assessment of empyema

History: the most important features of which will be useful to determine 1) why did this happen, 2) what is the most likely organism, and 3) the natural history.

• Risk factors for developing a post-pneumonia empyema:
  • Diabetes
  • Alcohol use
  • Drug addiction
  • Malnutrition
• Risk factors for a random empyema out of the blue:
  • Poor dentition
  • Aspiration pneumonia
  • Malignancy
  • Haemothorax, eg. following frequent falls or chest surgery (if left undrained, the risk is something like 26.8% in one study by DuBose et al, 2012)
• Recent hospital stay (hospital-acquired empyema is more likely to develop into a complicated loculated collection)

Examination focuses on:

• Completing the search for risk factors:
  • malnutrition
  • sequelae of IV drug use or alcoholism
  • end organ damage of diabetes
• Complicating comorbidities:
  • Cardiac failure
  • COPD

Investigations:

• Standard bloods, focusing on inflammatory markers
• Bedside ultrasound to image the extent of the collection and assess bedside drainability
• CXR  to assess the rest of the lung
• CT chest  with IV contrast is the gold standard for imaging an empyema, as it may demonstrate the enhancement of the "rind" 
• Pleural aspirate and culture is the definitive method of making the diagnosis, with the pleural fluid tested for:
  • Protein
  • Cell count
  • pH
  • glucose
  • cytology

Options for management

"List your options for the drainage of the empyema including one advantage and one disadvantage each", asked Question 4 from the second paper of 2022. This assumes you have settled on drainage as the solution. On a fundamental level,. that sentiment is correct (draining works best) but conservative management is often resorted to, as the patient may not yet be fit for anything more aggressive. For the purposes of converting this section into the answer, the content is offered in the form of a table, but expanded to include some of the conservative options, for completeness. The main sources for the advantages and disadvantages here were Light (2006), UpToDate, AATS and BTS.

Options for Draining (or Not Draining) an Empyema

Options	Advantages	Disadvantages
Uncomplicated parapneumonic effusion
Antibiotics alone	Usually enough (and the patient is already on antibiotics anyway for the pneumonia) Very few will go on to require drainage (~10%)	Effusion may become loculated with time (perhaps 48 hrs), making future drainage more difficult
Thoracocentesis	Diagnostic Easily repeatable	May recollect
Fine bore drain	May improve respiratory performance; May prevent infection of the pleural space	Not necessary in a large number of cases, and carries a risk of introducing infection into a previously sterile space
Empyema
Fine bore drain	Easily done at the bedside Better tolerated than wide bore drains Could be definitive in  a large number of cases (78% in one series)	Easily blocked Easily displaced Could delay other definitive management
Wide bore drain	Good rates of source control Considered the gold standard for viscous  fibrinous effusions	Not much benefit, when compared to fine bore tubes  Substantially more uncomfortable
Intrapleural fibrinolytics	Markedly improves rates of successful drainage, especially where loculations are beginning to form May reduce need for surgical decortication	Risk of haemothorax Inconsistent evidence of benefit (MIST1)
Loculated organising empyema
IR-guided drainage	The locules may communicate and late stage effusions may be thinner and less viscous, i.e. fine bore drains may still be effective	Unlikely to achieve source control if there are many loculations Multiple drains may be required
Thoracoscopic drainage	Breaks adhesions and allows drainage and correct tube placement Shorter recovery time than decortication Can be upgraded to decortication if a thick rind is discovered	The patient has to be well enough for one-lung ventilation Greater risk of complications, because more invasive
Decortication	Removes all pleural infected tissue and allows the lung to reexpand Viewed as definitive	Relatively long recovery time (~ 7days) Pain control is a major issue post-op Large incision; unsuitable to patients with poor functional baseline
Open drainage (short tube + bag, or gauze packing)	Frees the patient from the suction system Allows outpatient management in some cases Eventually creates a skin-lined fistula, allowing drainage to complete slowly	This is in effect an open chest wound May cause pneumothorax Depending on technique, may require repair in the future

It probably goes without saying that antibiotic therapy is essential irrespective of which drainage option one picks, and the duration of treatment increases by 1-2 weeks with each evolving stage of empyema, such that the uncomplicated parapneumonic effusion scarcely needs more antibiotics than the pneumonia that caused it, whereas a loculated organising empyema requires at least six weeks if not more. And the antibiotics may need to be different, as not all of them are equally well equipped to perform their role in the devascularised acidic cavity that's formed in your pleural space. 

Which brings us to:

Intrapleural fibrinolysis

If the septae are mostly fibrin, then it stands to reason that fibrinolytic agents should be able to destroy them. This is the rationale for injecting agents typically reserved for IV thrombolysis into the pleural cavity. DNAse (dornase alfa) was also added in the MIST2 trial, with some apparent benefit (radiologically looking better, and fewer surgical referrals). The doses were 5 mg of dornase and 10mg of alteplase, given twice daily via the (clearly fine-bore) ICC for three consecutive days. The outcome from this and several other trials had demonstrated a trend towards faster resolution and less frequent need for surgery. 

Factors that influence antibiotic choice

The antibiotic answers for CICM exam questions would most likely be derived from the Australian eTG (Therapeutic Guidelines) which means the standard choice of antibiotics for a bog-standard empyema would have to be:

• For uncomplicated small parapneumonic effusions, the standard antibiotics for pneumonia are enough
• For empyema complicating community-acquired pneumonia, TG recommend benzylpenicillin or ceftriaxone + metronidazole, or moxifloxacin solo.
• For empyema resulting from hospital-acquired or ventilator-associated pneumonia, TG suggest piperacillin/tazobactam as a sole agent, or cefepime and metronidazole if allergic. 

But those are the usual community-acquired streptococci and staphylococci, or the random Kleb you might grow in the ICU. Question 4 from the second paper of 2022 had MRSA, which changes the antibiotic choices substantially. Specifically, the college asked about the advantages and disadvantages of vancomycin and linezolid for this indication. Theoretically other agents (teicoplanin, daptomycin, tigecycline, ceftaroline) could also be thrown in to the mix, which means future candidates may get some benefit from tabulating these as well, but for now this section will focus only on what the college asked:

• Vancomycin:
  • Advantages:
    • Pharmacologically inert, no drug interactions
    • Cheaper
  • Disadvantages:
    • Narrow therapeutic index; requires monitoring
    • Requires IV access, ideally central venous
    • May be nephrotoxic
    • Requires adjustment in renal failure
    • Does not penetrate especially well into abscess cavities or lungs
• Linezolid:
  • Advantages:
    • No need for monitoring
    • Good oral bioavailability (i.e. an extended course can be given orally after the IV period is finished)
    • Does not need adjustment for poor renal function, nor is it nephrotoxic
    • Penetrates well into abssesses and lungs
    • Superior to vancomycin in RCTs (eg. Wunderink et al, 2003)
  • Disadvantages:
    • Has substantial drug interactions (eg. can cause serotonin syndrome)
    • More expensive as an IV formulation
    • Can produce thrombocytopenia

Why not infuse the antibiotics directly into the cavity

Why not indeed, the reader may wonder, confused by the appealing simplicity of the proposal. The bugs are in there, so why not also send the antibiotics directly into the cavity, maximising their effect site concentration? What could possibly go wrong. Indeed, there is some data that this is at least safe, if ot effective. Torbic et al (2015) reflected in retrospect on their experience of using the same IV agents intrapleurally in lung cancer patients, and Torbic et al (2008) presented a prospective randomised study where half the patients got neomycin and bacitracin twice a day directly into the chest cavity.

The technique, from both studies:

• Use antibiotics of a similar dose and concentration as one might use intravenously, so as to not inadvertently overdose the patient; and if persisiting with IV antibiotics, at least ensure that the total dose is within the normally accepted range 
• As an example, vancomycin meropenem voriconazole and ceftiazidime were all reported in Torbic et al (2015), who infused them as 1g in 1000ml NaCl 0.9% over 1-2 hours and in divided doses
• Not many doses are required. The median number of doses was 5.5 for Torbic et al, and only 72% of patients had more than a single dose. On the other hand, Ader et al gave their mixture for 8 days.
• Systemic absorption occurs to the point where one would wish to continue monitoring drug levels
• The presence of cavitating lesions and any communication between the pleura and the bronchi (eg. bronchopleural fistula) will result in not only drowning the patient in infusion solution (with resulting cough and shortness of breath) but also potentially the wash of infected necrotic debris into the healthy lung. 

However:

• Pretty much everything other than aminoglycosides penetrates reasonably well into the pleural cavity (Taryle et al, 1981, found pleural concentrations were mostly ~ 75% of plasma concentrations)
• The studies did not demonstrate much patient-centred benefit, and they were small and mostly retrospective
• Most of the data comes from patients whose chest cavities have already been cleared surgically, and in whom the intrapleural irrigation was mostly a recurrence prevention measure