Pneumocystis Jirovecii Pneumonia in untreated HIV infection

Thanks to Dr. Michael Chung for presenting today’s case about a young transgender woman with a history of untreated HIV who presented with gradual onset fever, cough, and acute hypoxic respiratory failure with a high A-a gradient, found to have PJP pneumonia on induced sputum. A special thanks to Dr. Kirsch for attending morning report and giving us stellar clinical pearls.

The differential for pulmonary symptoms in an HIV patient is broad, including typical viral, bacterial, and fungal etiologies with the addition of PJP, TB, MAC, toxo (other parasites including strongyloides, microsporidium, and crypto), lymphoma, Kaposi’s sarcoma, and immune reconstitution syndrome (if ART has been recently started).


Basics: HIV and severely immunocompromised

PJP is the most common AIDS-defining opportunistic infection in the US. It affects severely immunocompromised patients including HIV with CD4<200, hematologic malignancies, transplant patients, chronic steroids, and severe malnutrition. Risk factors that should raise your clinical suspicion for PJP are history of previous PJP or recurrent bacterial pneumonia, high viral loads, unintentional weight loss, or presence of oral thrush.

Clinical Presentation:

Basics: 3 weeks of nonproductive cough, dyspnea, and constitutional symptoms, now with hypoxic respiratory failure

HIV patients with PJP typically present with gradual worsening of symptoms over days to weeks (average 3 weeks): nonproductive cough (95%), fever (80-100%), and dyspnea (95%). Chills, chest pain, and unintentional weight loss are also commonly present. Physical exam usually reveals tachypnea and desaturations, but lung exam can be clear in 50% of patients.

Lab Evaluation:

Basics: LDH and B-D Glucan are sensitive, but nonspecific indicators of PJP that can help you decide to treat empirically if confirming the diagnosis in a timely manner is impractical. 

LDH is nonspecific, but very sensitive (elevated in >90% of HIV patients with PJP). LDH has been proposed as a prognostic marker in PJP pneumonia as those who do not survive PJP despite treatment were observed to have higher initial LDH levels than those who did survive. LDH can also be trended throughout treatment to evaluate response and prognosis. B-D-glucan, a component of fungal cell walls, is also nonspecific, but is elevated in PJP pneumonia. Due to the high mortality of untreated PJP pneumonia in HIV patients, these lab tests can raise your clinical suspicion for the disease and help you decide to treat empirically in the absence of confirmed diagnosis.

It is necessary to get an ABG in all patients with suspected PJP pneumonia to calculate the Alveolar-arterial oxygen gradient. This indicates severity of disease and will be useful in deciding on whether to give steroids with your antibiotics.

A-a gradient at sea level on room air= [150- PaCO2/0.8] – PaO2. Normal is <10 but increases with age.


Basics: PJP can cause almost any presentation on CXR/CT but diffuse bilateral infiltrates is most common. CXR is not sensitive so get a CT if clinical suspicion is high. 

25% of chest XRs of HIV patients with PJP are normal. The most common manifestations is diffuse bilateral infiltrates, but PJP can present in almost any way. CT scan is 100% sensitive for PJP in HIV patients.


Basics: Don’t delay treatment for diagnosis. Induced sputum first, then BAL if negative. 

If clinical suspicion is high, diagnosis should not delay treatment. Treat empirically in these cases. Diagnosis can be made with sputum induction (variable sensitivity 50s%-90% depending on patient characteristics, sputum characteristics, and skill of RT/lab/pathologist). If negative, proceed to bronchoscopy with BAL (sensitivity up to 90%). If clinical suspicion is extremely high, but these tests are negative, lung biopsy with PCR can be done, but is very rarely indicated and very invasive. Usually it is only done when lung biopsy is needed to diagnose other pathology concurrently. Silver stain with cysts or trophozoites is the typical method of detection for board exams.

At SCVMC, induced sputum to diagnose PJP requires a pulmonology consult due to the specialized nature of the induction procedure and patient selection and possibility of bronchoscopy.


Basics: Treat with Bactrim. Use adjunctive steroids if significant hypoxia. Things will get worse before they get better. 

Treat with Bactrim 15-20mg/kg PO or IV (equal bioavailability) x 21 days. If sulfa allergy, desensitize in the MICU and treat. If SJS or TEN with sulfa drugs, consider alternative regimens (clinda + primaquine, trimethoprim + dapsone, atovaquone, pentamidine). Always check for G6PD deficiency before starting primaquine or dapsone. Keep in mind that pentamidine is rarely used due to severe side effects: hypotension, hypoglycemia, nephrotoxicity, pancreatitis. It is often necessary to monitor these patients in the ICU for treatment due to the propensity of respiratory failure to initially worsen with treatment before getting better.

Give adjunctive steroids if PaO2 <= 70 or A-a gradient >= 35 or hypoxemia on pulse ox. Steroids have been shown to decrease mortality in these patients without an increase in other opportunistic infections.


Key Points:

  • HIV (CD4<200) and severely immunocompromised get PJP
  • Presents with 3 weeks of nonproductive cough, dyspnea, and constitutional symptoms, now with hypoxic respiratory failure
  • LDH and B-D Glucan are sensitive, but nonspecific indicators of PJP that can help you decide to treat empirically if confirming the diagnosis in a timely manner is impractica
  • PJP can cause almost any presentation on CXR/CT but diffuse bilateral infiltrates is most common. CXR is not sensitive so get a CT if clinical suspicion is high. 
  • Don’t delay treatment for diagnosis. Induced sputum first, then BAL if negative. 
  • Treat with Bactrim. Use adjunctive steroids if significant hypoxia. Things will get worse before they get better.

HTG-Panc – Don’t Pan(i)c!

Today we discussed a case of hypertriglyceridemia-induced pancreatitis (TG >4k). We first utilized this case as an opportunity to provide a comprehensive overview of acute and chronic pancreatitis:


We then discussed the management strategy:

General Pancreatitis Management

  1. Assess severity of pancreatitis with SIRS
  • Mild – absence of organ failure and local systemic complications
  • moderate – local complications and +/- transient organ failure (<48hrs)
  • Severe – persistent organ failure (>48hrs)
  1. Aggressive hydration at 5-10mls/kg/hr
  • Adjust based on vital signs, exam, urine output BUN, HCT
  1. Pain Control
  2. Nutrition
  • Mild
    • initiate PO intake within 48 hours (based on symptoms, NOT lipase reduction)
    • Go straight to low residue, low fat, soft diet
  • Moderate – Severe
    • If unable to tolerate PO, consider NJ vs. NG tube after 3-4 days
  • Enteral nutrition >>> parenteral nutrition
  1. Antibiotics?
  • not recommended regardless of pancreatitis severity or type (i.e interstitial or necrotitizing) unless extrapancreatic infection suspected (blood stream, PNA, UTI etc.) OR necrotizing fluid collection suspected to be secondarily infected (systemic infection OR radiographic evidence of infection)

HTG-Pancreatitis Management

  • Insulin Therapy
  • Lipid Lowering Therapy
  • Plasmapheresis (controversial)
    • invasive
    • expensive
    • requires AC
    • Despite that – UTD recommends apheresis in patients with any of the following (hypocalcemia, lactic acidosis, two or more signs of worsening inflammation, worsening organ dysfunction)
    • Two studies of note that discuss this
      • 1) Head-to-head RCT showed quicker time to TG reduction compared to insulin but worse clinical outcomes
      • 2) Addition of apheresis to patients already receiving insulin showed no clinically significant differences in outcome
    • Bottom Line: based on the limited data above, some experts favor an insulin only approach
    • For an in-depth analysis of this topic – please see this fantastic PulmCrit blog post:

Gout Attack!!!!

Today’s case was a 70yo man from the Philippines with active colorectal cancer on chemotherapy as well as a history of gout and CKD4 who presented with a week of subjective fevers and knee pain and swelling refractory to his normal outpatient gout management.

Clinical Pearls:

  • Gout and septic arthritis can present very similarly. In an immunocompromised patient, always consider septic arthritis even if the pain is not severe.
  • Triggers for gout flares include dietary noncompliance, alcohol, med noncompliance, diuretics, and chemotherapy
  • Tuberculosis can cause monoarticular septic arthritis, but the most common form of skeletal TB is Pott’s disease (spinal TB). Extraaxial skeletal TB affects large joints usually (hip or knee) and can be a “cold” joint without erythema or warmth.
  • Interpretation of synovial fluid studies:
    • WBC <2000 noninflammatory (trauma, OA, etc)
    • WBC >2000 inflammatory vs. crystalline vs. septic
    • The higher your WBC count, the more concerned you should be for septic arthritis
    • Gram stain and culture all synovial fluid to evaluate for septic arthritis, which can occur simultaneously with a gout/pseudogout flare or inflammatory arthropathy.
    • Gout: Negatively birefringent crystals (yellow, needle-shaped)
    • CPPD: Positive biregringent crystals (blue, rhomboid)


For a comprehensive review of gout presentation, diagnosis, and management, see the following blog post: Or search gout on the right side of this page.

CD4<10 – A Major Headache

Today we reviewed a case of a patient with HIV and a last known CD4<10 who is intermittently on ARVs who presented with 2 weeks of a headache, neck stiffness and photobia

We first reviewed the general framework for headaches

  • Primary
    • migraine
    • tension
    • cluster
  • Secondary
    • Intracranial
      • Vascular
        • stroke
        • aneurysm
        • vasculitis
        • systemic HTN
        • hemorrhage
        • dissection
      • Infection
        • meningitis
        • encephalitis
        • abscess
        • septic emboli
        • toxoplasmosis
      • Tumor
        • metastasis
        • primary
      • CSF
        • pseudotumor cerebri
        • hydrocephalus
        • low-pressure (post-LP or leak)
      • Other
        • trigeminal neuralgia
        • post-traumatic
    • Extracranial
      • meds
      • alcohol withdrawal
      • TMJ
      • dental abscess
      • cervicogenic
      • GCA

We then discussed the specific bugs you must consider in a patient with HIV presenting with a headache

  • Cryptococcus
  • TB
  • Toxoplasmosis
  • HSV infection
  • Coccidiomycosis
  • CNS primary lymphoma (not a bug but a unique condition that must be considered in an HIV patient with a headache

CT Before LP? The IDSA Describes Six Scenarios Where You Must Do It

  • Immunocompromised state
  • History of CNS disease (eg, mass lesion, stroke, or focal infection)
  • Seizure within 1 week of presentation
  • Papilledema
  • Abnormal level of consciousness
  • Focal neurologic deficit (eg, dilated nonreactive pupil, gaze palsy, or arm or leg drift)

Empiric Antimicrobial Therapy In Suspected Meningitis in Immunocompromised Patients

  • Vancomycin/Cefepime/Ampicillin OR Vancomycin Meropenem (you must cover for pseudomonas and listeria in immunocompromised patients regardless of age or other risk factors
  • All immunocompromised patients should be covered empirically with Acyclovir for HSV

Our patient did not have Cryptococcal Meningitis, but his clinical presentation is classic for it

  • The cryptococcal antigen (serum or CSF) is 95% sensitive and specific for the presence of cryptococcus (AKA a really good test). If you don’t find it in the CSF, it’s not crypto meningitis!
  • India Ink is at best 80% sensitive (expertise + high fungal load needed for direct visualization)
  • If it were cryptococcal meningitis, treatment is as follows:
    • Induction:Ampho B + flucyotosine for minimum of 2 weeks.
    • Consolidation:Fluconazole 200-400mg daily x 8 weeks
    • Maintenancetherapy is continued for at least 1 year.
    • Intracranial pressure control:Cryptococcal meningoencephalitis patients might need daily LPs to relief ICP, might require VP shunt. Do not use mannitol or acetazolamide.
    • Steroids not shown to be helpful.
  • In cryptococcal meningitis, wait 4-6 weeks to initiate ARVs to prevent the risk of IRIS

A Highly Provocative Seizure

Today we discussed a case of a patient found to have focal, motor seizure symptoms who was eventually found to have a frontal tumor with vasogenic egema that is likely a metastases from a lung malignancy.

We first defined the following:

Seizuresudden change in behavior caused by electrical hypersynchronization of neuronal networks in the cerebral cortex

There are two aspects of seizure classification that are important to consider: the cause/setting of the seizure and then the location/type of seizure.

Cause/Setting of Seizure

  • Acute Symptomatic – seizure that occurs in close proximity with a brain or systemic insult (25-30% of firs seizures)
    • Brain
      • Subdural hematoma
      • SAH
      • TBI
      • Hypoxic-ischemic injury
      • Brain abscesses
      • Meningitis/encephalitis
    • Systemic
      • Alcohol Withdrawal
      • Drug Intoxication
      • Hyponatremia/Hypernatremia
      • Hypomagnesemia
      • Hypocalcemia
      • Hypoglycemia
      • Nonketotic hyperglycemia
      • Uremia
      • Hyperthyroidism
      • Dialysis Disequilibrium Syndrome
      • Porphyria
  • Unprovoked – unknown etiology OR one that occurs due to a preexisting brain lesion or progressive nervous system disorder. the latter is referred to as remote symptomatic seizures.
    • medial temporal lobe epilepsy
    • Poststroke
    • Primary or metastatic brain tumors
    • Vascular malformations
    • Prior CNS infection
    • Head injury
    • Neurodegenerative dementia

Types/Location of Seizure

We reviewed the revised 2017 International League Against Epilepsy Classification:

classification of seizures.PNG

When Do We Get Neuroimaging?

  • Short answer: ALL patients with a first time seizure should have neuroimaging to evaluate for a culprit structural abnormality (some guidelines present some rare exceptions but this should be deferred to the experts)
  • MRI is preferred due to superior sensitivity and results sometimes need to be interpreted with caution as some findings may be non-specific and not related to the index event
  • In urgent cases start with a head CT and then in most cases an MRI (i.e an intracranial lesion is suspected OR new focal deficits OR persistently altered mental state, fever, persistent headache, focal-onset seizure, history of head trauma, malignancy, immunocompromise, alcoholism, anticoagulation or bleeding diathesis)
  • Per UTD, deferred outpatient MRI from the ED is mentioned in patients who have a normal neurologic examination but only if they have a normal CT And reliable follow up can be ensured (systemic limitations on MRI scheduling may not make this recommendation feasible for every situation

When Should We Start Antiepileptics?

  • Short answer: it depends
  • Acute symptomatic seizures in the critically ill
    • IV AEDs should be started to prevent recurrent seizures and prevent further destabilization
  • Acute symptomatic seizures provoked by metabolic derangements
    • they may still be at risk for short term seizure recurrence particularly if the metabolic process may take time to resolve – therefore short-term AED use could be considered
  • First unprovoked seizure with a full return to baseline
    • this is the controversial scenario
    • 1/3 of these patients will have a recurrent seizure within 5 years and the risk is increased 2-2.5x in any of the following factors
      • Epileptiform abnormalities on interictal EEG
      • Remote symptomatic case by history or imaging (i.e structural brain abnormality)
      • Abnormal neurologic examination, including focal findings or intellectual disability
      • A first seizure that occurs during sleep
    • In patients with the above risk factors, initiation of AEDs suggested in most patients
    • In patients without the above risk factors, care is individualized and specific to the risks/benefits in your specific patient

Acute Painful Monocular Vision Loss… It’s simple! No really… it is

Today we discussed a case of a 38 year old man with acute onset left eye vision loss with pain on eye movements and without eye redness. For an internist, the chief complaint of vision loss is often intimidating, but when we examine the facts with an anatomic framework, it becomes surprisingly simple.

Below is a diagram that includes our differential for acute, persistent, monocular vision loss. For completeness, retrochiasmal anatomy is included, but pathology occurring after the optic chiasm (in the optic tracts, Meyer loop, lateral geniculate nucleus, optic radiations) will cause bilateral homonymous hemianopsia, not monocular vision loss. Blindness occurring in the occipital lobe is termed cortical blindness and is bilateral and severe.

Capture When we consider pain with eye movements in our illness script, our differential narrows considerably to the following entitities:Capture

Note that papilledema and GCA are not included in this chart because they include pain due to headaches, but not pain with eye movement.

Finally, when we consider that our patient’s eye was not injected, erythematous, or tearing, our differential narrows still further to endopthalmitis and optic neuritis. On fundoscopic exam, the clinical diagnosis of optic neuritis can be made if papillitis is seen without hypopyon, WBCs in the vitreous humor, or white mound-like lesions in the retina.

A Dizzying Topic

Today we discussed a case of a patient who had been reporting 6 months of “dizziness”. We first took a deep dive into dizziness and what exactly it means.

The single most important question to ask a patient with dizziness is: what do you mean by that?

Dizziness can then be classified into four major categories depending on the description of the symptoms (though of course this slightly antiquated framework does have limitations). We emphasized the importance of not coaching the patient to using certain descriptors and also recognized the cultural differences in the usage of the term dizziness (many communities utilize this word to indicate sickness, rather than how we define it).


We focused our discussion on Vertigo as the patient reported a room-spinning sensation for the past 6 months. The major discussion is always central vs. peripheral vertigo and please see the excellent prior blog post discussing this in detail:

We reviewed some important distinguishing characteristics (as taken from a 2014 NEJM Review Article on BPPV)


The key takeaways are as follows:

1. BPPV is characterized by positional, brief (<1 minute) attacks of severe vertigo prompted by head movements and if nystagmus is seen, it is unidrectional. The typical discussion therefore is not BPPV vs a posterior CVA but rather an acute vestibular syndrome (vestibular neuritis/labrynthitis) vs. a posterior CVA.

2. If the Dix-Hallpike induces vertigo, this does not mean the patient has BPPV! The characteristics of the nystagmus however may be very helpful


3.  The HiNTs exam can be exceptionally useful, but only in the right clinical context (it can only be used if the patient is currently symptomatic, and is meant for continuous, not intermittent, symptoms. (i.e hours – days)

  • Recognize the difficulty of performing the examination well (the initial study was done by one single neuro-opthalmologist and as far as we know, no inter-reliability study has been performed with a group of generalists
  • Alarming findings are useful though and should prompt further investigation
  • Remember, ALL three of the following components must be seen in order to state that the exam is truly negative (and therefore a peripheral vertigo)
    • Unidirectional or no nystagmus
    • No vertical skew
    • Abnormal head impulse test (corrective saccade) – this indicates a nerve problem and therefore must be seen in order to state that the HiNTs exam is negative and therefore the vertigo is peripheral

4.  Try not to overly rely on imaging, especially a head CT, which is at best, 16% sensitive for acute ischemia in the posterior fossa. Even MRI with DWI misses 15-20% of acute posterior fossa infarctions in the first 24 hours. Sensitivity for brainstem strokes is maximal 72-100 hrs after infarction