Definition, causes and differential diagnosis of ARDS

"What if ARDS, but in the [insert patient context]" is a common trope of CICM exams, where the exercise of generating a broad range of differentials and a list of clever investigations in under ten minutes closely resembles what one ends up doing at the bedside. Well, of course, it is also possible to completely decompensate when faced with an unfamiliar problem, turning into a consult geyser, but this is not the standard we intend to model for our trainees, and so the college examiners has prioritised this aspect of our practice. It is usually tagged as "Respiratory Failure" (an  L1 condition) plus whatever else (eg. "Lung Transplantation", and L2 sidequest from Section 2.1.5 in the second edition of the CICM Syllabus for the Second Part Examination), but it could legitimately be grouped with sepsis (its often infectious) or under radiology, as the finding of "diffuse infiltrates" is difficult to appreciate from any other assessment modality.

Many past paper questions ask about the causes and differential diagnosis of " a diffuse bilateral infiltrate on CXR." There are many scenarios available.  It is important to be able to generate a lot of differentials in this sort of SAQ.

• Question 13 from the second paper of 2024: 
  what if inflitrates, but in the lung transplant recipient?
• Question 14 from the first paper of 2021:
  what if inflitrates, but on chronic steroids for RA?
• Question 4 from the second paper of 201:
  what if inflitrates, but it's an APML patient receiving ATRA?
• Question 20 from the first  paper of 2014:
  what if infiltrates, but on chronic steroids for RA?
• Question 17 from the second paper of 2011:
  what if infiltrates, but it's a bone marrow transplant recipient?
• Question 10 from the second paper of 2010:
  what if infiltrates, but after a recent cardiac arrest?

Definitions of ARDS

The most recent definition (Matthay et al, 2023) presents a comprehensive model. The contents of their Table 1 is presented below with minimal structural modification:

Definition:

• ARDS is an acute diffuse, inflammatory lung injury precipitated by a predisposing risk factor such as pneumonia, non-pulmonary infection, trauma, transfusion, burn, aspiration, or shock.
• The resulting injury leads to increased pulmonary vascular and epithelial permeability, lung edema, and gravitydependent atelectasis, all of which contribute to loss of aerated lung tissue.
• The clinical hallmarks are arterial hypoxemia and diffuse radiographic opacities associated with increased shunting, increased alveolar dead space, and decreased lung compliance.
• The clinical presentation is influenced by medical management (position, sedation, paralysis, and fluid balance).
• Histological findings vary and may include intra-alveolar edema, inflammation, hyaline membrane formation, and alveolar hemorrhage.

Criteria

• A predisposing risk factor:
  • Eg. pneumonia, non-pulmonary infection, trauma, transfusion, aspiration, or shock
  • Pulmonary edema is not exclusively or primarily attributable to cardiogenic pulmonary edema/fluid overload
  • Hypoxemia/gas exchange abnormalities are not primarily attributable to atelectasis
  • However, ARDS can be diagnosed in the presence of these conditions if a predisposing risk factor for ARDS is also present.
• Acute onset: within 1 week of exposure to the risk factor
• Radiological criteria:
  • Bilateral opacities on chest radiograph and computed tomography, OR:
  • Bilateral B lines and/or consolidations by ultrasound
  • Not fully explained by effusions, atelectasis, or nodules/masses.
• Oxygenation criteria:
  • If in a resource-limited setting: 
    • SpO₂/FiO₂ less than 315 (if SpO₂ at least lower than 97%)
  • If not intubated: 
    • PaO₂/FiO₂ less than 300 mmHg, OR
    • SpO₂/FiO₂ less than 315 (if SpO₂ at least lower than 97%)
    • on HFNO with a flow of at least 30 liters per minute, OR
    • on NIV/CPAP with at least 5 cm H2O expiratory pressure
  • If intubated: 
    • PaO₂/FiO₂ ratio less than 300 on a minium PEEP of 5

      ARDS Severity	PaO2/FiO2	SpO2/FiO2	Mortality
      Mild	200 – 300	235 – 315	27%
      Moderate	100 – 200	148 – 235	32%
      Severe	< 100	< 148	45%

This is a revision of the 2012 Berlin definition, where the specific changes were:

• Adding the idea that the onset may be more indolent (into the acute onset criteria)
• Adding ultrasound into the radiological criteria (because not everywhere has Xray facilities available)
• Adding pulse oximetry criteria (because these are well validated)
• Adding a whole set of categories for different resource settings and incorporating patients who are not intubated

The Berlin definition in turn improved on even older criteria. The changes made in 2012 were:

• There is no longer such a thing as "acute lung injury" when you have a PaO₂/FiO₂ ratio of 200-300.
  We call it "mild ARDS" instead.
• Onset must be acute: within 7 days of whatever pathology is thought to be the cause
• There is no need to exclude heart failure - you can have as much heart failure as you like, so long as its severity is insufficient to by itself explain the bilateral pulmonary infiltrates.

Differential diagnosis of pulmonary infiltrate in ICU patients

An excellent article from Silvia Blanco and Antoni Torres (antimicrobe.org) actually contains a brilliant table of differentials, which is incorporated into the table below. 
 

Idiopathic pneumonia syndrome and ATRA syndrome(nowadays referred to as "differentiation syndrome") have been added since Question 4 from the second paper of 2015 introduced them into the list of differentials. This table is otherwise rather generic. A lot of it would be irrelevant to the pregnant patient, the patient recently fished out of a pond, a patient recently undergoing thirty minutes of CPR, or a patient with a massively underperforming immune system. The latter seems to be a much more common type of CICM SAQ, and so a specific list of causes is presented here:

And for the recipient of the lung transplant, who find their way into CICM exam questions quite often, several special mentions need to be listed:

• Acute rejection
• Chronic Lung Allograft Dysfunction
• Bronchial anastomosis stenosis/ breakdown with underlying collapse of the lung segments
• Neoplasm rises higher in an ordered list of likelihood: cancer is the second most common cause of death in the lung transplanted population (Shtraichman & Ahya, 2020)

Drug-induced pneumonitis  is an unpleasantness inflicted on the immunosuppressed by the very drugs that suppress them, as there is a group of immunosuppressant agents with known pulmonary toxicity in addition to all the usual suspects that cause drug induced pneumonitis. These are listed by Meyer (2014):

• Methotrexate (causes an NSIP)
• Sirolimus (also a weird NSIP
• Cyclophosphamide (causes a random idiosyncratic AIP, with diffuse alveolar damage)
• Less frequently, azathioprine, leflunomide, mycophenolate

Assessment for unexplained ARDS

The usual assessment process needs to answer the following questions:

• Is this an infection
• What kind of infection is this
• How is this not an infection

The main reason for these comical tautologies is that the two main weapons in our therapeutic armamentarium are antibiotics and megadose steroids, and so it makes logical sense to exclude infection before ablating the immune system with a pulse of methylpred. Ergo, the process of investigating ARDS involves the exploration of the infectious possibilities first. Papazian et al (2016) gives a solid foundation for this assessment process, though most readers at the final stages of their CICM training will not be impressed by the breadth of the differentials or surprised by the investigations suggested.

History for new unexplained ARDS needs to focus on:

• Immunocompetence (as this determines the breadth of investigations)
  • Transplantation, chemotherapy, autoimmune disease, HIV
  • any possibility that the patient may be functionally or actually anatomically asplenic
  • Chronic conditions that predispose to lung disease (eg. cystic fibrosis, alveolar proteinosis)
• Infectious possibilities
  • Recent travel
  • Suspicious occupational or recreational history (birds?  dust? livestock? gardening? etc) 
  • Exposure to others with similar illness?
  • High risk background (eg. from a TB-endemic country, or poorly controlled diabetic, heavy smoker, alcoholic, etc)
• Noninfectious possibilities
  • Rash, join swelling, haematuria (suspicion of vasculitis)
  • Slow, long course
  • History of ineffective courses of antibiotics
  • Weight loss (malignancy) or gain (cardiac failure)
  • Spent some time recently without airway reflexes, for reasons? (aspiration)
• For the immunosuppressed patient, also:
  • History of immunosuppressant use and adherence
    • Levels?
  • Empirical prophylaxis use
  • Vaccinations (splenectomy?)
  • Recent drug changes (interactions?)
  • Intestinal unhappiness that might have resulted in the malabsorption of oral antirejection drugs

Examination in the new workup of ARDS calls for:

• Finding features of infectious illness:
  • Peripheral features of infective endocarditis
  • Pharyngitis
  • Purulent sputum
  • Shocked state, vasodilated
• Finding features of noninfectious illness
  • Multiorgan involvement
  • Characteristic eye, skin or joint/musculoskeletal findings (scleorderma, RA, CREST syndrome, dermatomyositis, ankylosing spondylitis)
  • Extrapulmonary sources of inflammation, eg. pancreatitis, ischaemic limb, sore muscle compartments (polymyositis)
• Finding such as fever, rash, lymphadenopathy, splenomegaly, as well as ausculation findings, are sufficiently scattered across the diagnostic categories as to be entirely meaningless from a diagnostic standpoint
• In the immuno-abnormal host, 
  • Haematuria, haemoptysis, jaundice are suggestive of a more widespread MAHA and haemolysis, so hypothetically HLH
  • Haemodynamic stability suggests inflammatory or autoimmune causes
  •  

Investigations would obviously include:

• Unimaginative bloods: FBC, EUC, CMP, LFT, but especially the inflammatory markers such as procalcitonin WCC and CRP
• Cultures of everything
• Serology:
  • for atypical pneumonia (though, admittedly, convalescent samples will need to be collected to be absolutely sure, and by that stage these will be of largely academic interest)
  • donor specific antibody levels (for the lung transplant recipient)
• Bronchoscopy for BAL culture, as well as to assess the appearance of bronchial mucosa, which can offer clues
• Sputum for cell count and cytology (what if it's swarming with eosinophils?)
• PCR of the BAL sample, looking for the usual suspects (viruses, mycoplasma, the "typical" atypical such as viruses, Chlamydia, Legionella, B.pertussis, mycoplasma etc)
• For the immunocompromised patient, PCR for specific organisms:
  • For the transplant recipient who's had their cell-mediated immunity disabled, it would be viruses (the usuals plus CMV, VZV, HSV, EBV), intracellular opportunists such as Nocardia, and fungi like PJP, Cryptococcus, Aspergillus, etc.
  • For the chronic rheumatological patient who is forever on B-cell modulators, or for the asplenic patient, you would screen for common CAP encapsulated bacteria, like Haemophilus, as well as tuberculosis, Candida and Nakaseomyces.
  • For the chronically hospitalised neutropenic patient,  Pseudomonas-like saprophytes  would be at the top of the list (Citrobacter, Serratia, Stenotrophomonas, that sort of thing)
• CXR but realistically also a CT chest
• TTE to rule out cardiac causes or vegetations
• Bronchial or lung biopsy