College Answer

a) Differential Diagnosis:
1. Bacterial/atypical Pneumonia
2. Opportunistic Infections:
a. Viral: Influenza/CMV/other Herpes Viruses
b. Fungal: Aspergillus/Cryptococcus
c. Other Organisms: PCP/PJP
3. Related to Rheumatoid Arthritis:
a. Methotrexate-induced pneumonitis
b. Rheumatoid Lung Disease

4. Acute cardiac failure e.g. secondary to valvular heart disease, ischaemic cardiomyopathy

b) Management issues:
1. Diagnostic investigations:
a. HRCT/CTPA
b. Bronchoscopy +/- lung biopsy (level of respiratory support may determine whether these investigations are possible)
c. Echo
2. Cease methotrexate
3. Steroids
a. Consider increasing the dose of steroid to cover "stress response"
b. Consider treatment dose associated with PCP/PJP treatment
c. Consider high-dose pulse of steroids
4. Empirical anti-infective treatment (complex decision, treatment may be associated with toxicity)
a. Broad spectrum antibiotic e.g. 3rd generation cephalosporin/aminoglycoside
b. Atypical cover
c. Oseltamivir
d. High dose Co-trimoxazole: monitor for Myelotoxicity
e. Gangcyclovir: monitor for Retinitis, Myelotoxicity
5. Specific treatment for cardiac disease

Discussion

Differential diagnosis for bilateral pulmonary infiltrates should at first glance fall into the "is it infection or is it heart failure" territory. That would be the sensible approach. However, instead here is a long list of differentials.

The college then asked for "specific management issues relating to diagnosis and treatment", which according to their model answer called for a salad of recommendations ranging from stopping immunosuppresive therapies ("cease methotrexate") to starting more immunosuppressive therapies ("consider high dose pulse of steroids").  It is of course difficult to assess the situation or determine what specific management or investigations are called for, even with the relatively extensive history given here. Unusually, the examiners have furnished us with some very relevant details:

• Relatively young patient with no past respiratory problems
• Severe hypoxia (must be, because she needed to be intubated)
• Cough and dyspnoea (whereas dyspnoea on its own would lead you down a slightly different path)
• Sub-acute onset, over 4 weeks
• History of fairly heavy-handed immunosuppression

The problem of non-specific lung disease appearing randomly and lethally among patients with poor immune systems is sufficiently common in the ICU, to the effect that multiple articles pop up if one searches for "diffuse pulmonary infiltrates in the immunocompromised". 

To exclude non-infectious causes:

• A transthoracic echo will inevitably be informative, but the finding of a poor systolic function is not going to exclude infectious aetiology (which could easily co-exist with heart failure, as a wet lung is the devil's playground)
• HRCT is suggested by the college, and this may give some information regarding the pattern of the disease, while not being particularly diagnostic (the CTPA would reveal emboli, but surely a gradual onset over four weeks does not particularly resemble the natural history of a PE)
• A "vasculitic screen", whatever that might be in the local parlance - mainly to exclude something like Wegener's or Goodpasture's syndromes (though these are made unlikely by the immunosuppression) 

To investigate infectious causes:

• Perform a bronchoscopy and send lavage specimens for multiple tests:
  • Bacterial cultures and gram stain
  • Acid-fast bacilli
  • Cell count (also looking for weird stuff like eosinophilic pneumonitis)
  • P.carinii PCR
  • Aspergillus PCR
  • Respiratory viral nucleic antigen tests, including CMV, HSV and VZV
  • Cryptococcal antigen
• Urinary antigens for Streptococcus and Legionella
• Atypical pneumonia serology, looking for antibodies to mycobacteria, Chlamydia, Coxiella, etc.

Reasonable steps to prevent deterioration:

• Cease methotrexate. The disease process progressed while the patient (presumably) continued to dutifully take her methotrexate and steroids; ergo it is less likely to be an autoimmune disease driven by B-cells and auto-antibodies.
• Be moderate with fluid resuscitation, as this has been demonstrated to have a negative impact in ARDS
• Ventilate the patient with lung-protective low tidal volumes, high PEEP and minimal driving pressures

Some empirical management to cover for the usual suspects:

• Cover P.carinii with therapeutic dose of sulfamethoxazole/trimethoprim; 
• Cover gram-positive and gram-negative organisms with something broad, as the stakes are high and there will always be opportunity to narrow the antibiotics once cultures are available. Some combination of meropenem azithromycin and either vancomycin or linezolid are recommended by various guideline-writers (eg. the parts of the 2016 IDSA guidelines which mention "critically ill patients")
• Antifungal therapy might become relevant if fungi are implicated by culture results
• Antiviral therapy (oseltamivir) is suggested by the college; the evidence in support of this is very weak, but proponents might argue that those trials don't enrol patients like this, that we have nothing to lose by giving it, and that observational studies (eg. Rodriguez et al, 2011) have suggested that there might be a benefit.

If things are not going as planned (i.e. it's a week down the track and the patient is not getting better), a lung biopsy might be indicated. Apparently, it often identifies steroid-responsive pathology (Gerard et al, 2018), in which case the college's suggestion (massive doses of methylprednisolone) becomes relevant.