8/22 (HUP): controversies in the world of submassive PE

Today we discussed a middle aged woman with factor V Leiden who came in with dyspnea and was found to have a monster (seriously) saddle PE. We had the pleasure of having Dr. Jay Giri talk to us today about the evolving evidence in submassive PE management and the PE response team.

There was way too much discussed to put in one post, but I’ll try to distill the essentials.

PE risk stratification

Screen Shot 2017-08-22 at 10.54.36 AM.png

  • Low risk PEs (those with none of the features above) can be treated with just anticoagulation (and may not even need admission, depending on the situation)
  • High risk (‘massive’) PEs should usually be treated with fibrinolytics because they will otherwise almost certainly progress to death; uncontrolled bleeding is the main contraindication
  • Intermediate risk (‘submassive’) PEs are where the opportunity is ripe to prevent morbidity and mortality

Options for treating intermediate risk PE

  • Anticoagulation + systemic thrombolysis
  • Catheter directed intervention (thrombolysis, vacuum aspiration)
  • (much less commonly) surgical embolectomy

While it’s well-accepted that thrombolysis is the best option for massive PEs, the use of thrombolytics for submassive PEs has been the subject of intense study. Several trials and meta-analyses have looked at this question, so I’ll highlight a few:

  1. PEITHO: anticoagulation + tenecteplase vs anticoagulation alone in patients with submassive PE. The study confirmed that fibrinolytics decreased the risk of hemodynamic decompensation
    • This study was not powered to detect differences in mortality
    • The bleeding rate was ~2% in the lytic group; in subgroup analyses a lot of that was driven by patients >75 (see plot below)

      Screen Shot 2017-08-22 at 11.50.39 AM.png
      Meyer et al. NEJM 2014
  2. This 2014 JAMA meta-analysis looked at prospective randomized trials of thrombolysis in >2000 patients with submassive or massive PE
    • Thrombolysis reduced all-cause mortality by 44% with a NNT of 59 (p=0.01), an effect also seen among the subgroup with submassive PE
    • Also note (table below) that the NNH is only 18 for major bleeding (but NNH = 176 in patients <75!)
Screen Shot 2017-08-22 at 10.29.47 AM.png
Chatterjee et al. JAMA 2014

Overall, clinicians who deal with this frequently find themselves in the position of having to decide which patients with submassive PE might benefit from thrombolysis. Right now it’s a decision made without clear guidance, but younger age and fewer comorbidities (and patient preference) are likely to play big roles.

A few other pearls from today’s talk 

  • Why do older patients do so much more poorly with lytics? Many possibilities: underlying amyloid angiopathy, prior strokes (whether recognized or not), possible old unrecognized ICHs
  • Lytic therapy is known to decrease PA systolic pressures much more rapidly than just anticoagulation, but there may be a ‘catch up’ phenomenon, such that months after the incident PE, PASPs are approximately equivalent in patients treated with lysis vs just anticoagulation; this suggests that A/C works, just slower
  • Transthoracic echo is decent at estimating RV and PA pressures (as compared to RHC), but that close concordance falls apart at very high PA pressures (approximately >60)
  • If someone codes and you suspect it’s due to a massive PE, you could consider giving them emergent lytics, but those people rarely survive given the attendant high bleeding risks with CPR, anoxic brain injury, etc. The better thing may be to crash them on to VA ECMO, which in some series confers a higher survival rate
  • Some (albeit shaky) data exists that lower doses of tPA (50mg instead of 100mg) may be equivalent, since the entirety of the cardiac output flows through the pulmonary vasculature
  • Catheter directed lysis theoretically has a lower risk of major bleeding since the tPA dose (typically 8-24mg) is much lower than systemic lytic doses, but that lower bleeding risk hasn’t actually been proven

Finally, consider activating the PE response team for any patient who has a submassive or massive PE; the attending or fellow on-call can always be found on Penn Medicine on Call (UPHS homepage).


  1. Meyer et al. Recent advances in the management of pulmonary embolism: focus on the critically ill patient. Ann Intensive Care 2016.
  2. Piazza et al. Management of submassive pulmonary embolism. Circulation 2010.
  3. Submassive PE: are we treating it backwards? Pulmcrit/EMCrit.


8/15 (HUP): alcoholic hepatitis

Late post thanks to boards! A big thanks to Dr. Vandana Khungar for helping us through the case of a young man who presented with jaundice and abdominal pain with a bilirubin of 51 and was ultimately diagnosed with acute alcoholic hepatitis (for which he got a liver transplant).

Key features of alcoholic hepatitis

  1. Even though we often think of alcoholic hepatitis as ‘acute’, its actually more often a manifestation of chronic heavy alcohol intake (>80g/day for a decade or more).

Ever wonder how much alcohol is in a ‘standard’ drink? Take a look!

Screen Shot 2017-08-15 at 8.50.33 AM.png

Remember that alcoholic hepatitis (AH) can be superimposed on underlying cirrhosis/chronic liver disease.

  • Patients can present with fever, tender hepatomegaly and leukocytosis; it can sometimes be difficult to distinguish from other causes of acute hepatitis, as well as sepsis/infection
  • Some series have shown that >50% of patients have abdominal bruits; other stigmata of liver disease (ascites, encephalopathy) may be present either with or without underlying liver disease
  • The AST:ALT ratio will usually be >2:1: thought to be due to hepatic deficiency of pyridoxal 5′-phosphate in alcoholics (a cofactor for the enzymatic activity of ALT); so the ratio is really a failure of ALT to increase appropriately

Risk stratification for alcoholic hepatitis

There are several risk scores used for EtOH hepatitis:

Screen Shot 2017-08-15 at 10.25.34 AM.png
Lucey. NEJM 2009.


Maddrey’s discriminant function: estimates EtOH severity; a score >32 is considered severe.
Glasgow score: used to estimate likelihood of benefit from steroids. Rarely used in clinical practice
Lille score: helps decide whether to stop or continue steroids after 1 week (ie which patients won’t benefit from steroids no matter what given the severity of their hepatitis)

How is alcoholic hepatitis managed?

  • Supportive care (fluids, antibiotics if needed, abstinence, nutritional support)
  • The big debate is over the role of prednisolone and pentoxifylline (a TNF-a inhibitor) in the treatment of alcoholic hepatitis. You can read the details of the STOPAH trial below (which compared the two), but the short answer is that it’s not totally clear that one is better than the other.
  • One major flaw of STOPAH may have been that in the groups randomized to steroids, the investigators continued prednisolone for 28 days instead of using the Lille score to stop steroids for those people that wouldn’t have benefited; in doing so, they may have driven up mortality due to infection in the steroid group and diminished any apparent benefit of prednisolone!
  • In clinical practice (at least at Penn), prednisolone is often started after 48 hours of negative infectious workup, as long as they don’t have renal failure (a population that was excluded in STOPAH, due at least in part to concerns about bleeding risk)

See reference #2 below for more information about the role of liver transplant in treating alcoholic hepatitis!


  1. Thursz M et al. Prednisolone or Pentoxifylline for Alcoholic Hepatitis (STOPAH). NEJM 2015.
  2. Saberi, B. et al. 2016. Current Management of Alcoholic Hepatitis and Future TherapiesJournal of Clinical and Translational Hepatology. 4, 2 (2016), 113–122.
  3. Lucey et al. 2009. Alcoholic Hepatitis. NEJM.

8/17 (HUP): diffuse GGOs and PCP

Thanks to Tom Franzon and Ben Manning for presenting two (different) patients who presented with subacute shortness of breath and were found to have diffuse groundglass opacities (GGOs) on chest CT and were ultimately diagnosed with HIV for the first time.

Along the way, we went through a diagnostic approach for diffuse ground glass opacities. Developing a differential for just ‘GGOs’ is a Herculean task, but creating a differential for diffuse GGOs (as opposed to scattered GGOs, particularly in association with some other consolidation) is actually much easier to do.

Screen Shot 2017-08-17 at 8.16.29 PM.png
Diffuse GGOs in a patient with methotrexate pneumonitis. AJR 2005.

Here’s a table from a paper by Penn radiologist/wizard Wally Miller:

Screen Shot 2017-08-17 at 3.51.50 PM.png
Miller and Shah, AJR 2005

This man was ultimately diagnosed with PCP pneumonia.

A few key points about diagnostic testing related to HIV and opportunistic infections:

  • Beta-D-glucan has excellent sensitivity (~95%) and specificity (~70-80%) for the diagnosis of PCP; in a patient who truly has PCP, BDG levels will likely be >400-500
  • CD4 counts can decrease with periods of acute/critical illness, so the CD4% may be a more stable marker to follow in someone with HIV, as it shouldn’t change all that much
    • An absolute CD4 count >500 corresponds to a CD4% of >30%

    • An absolute CD4 count between 200-500 corresponds to a CD4% of 15 to 30%

    • An absolute CD4 count <200 cells/microL corresponds to a CD4% of <15%

  • Most of the time, non-HIV related low CD4 counts don’t predispose to opportunistic infections, with the exception of idiopathic CD4+ lymphopenia— patients with this condition behave similarly to HIV+ patients with low CD4 counts in terms of OI susceptibility, except that they don’t have HIV!
  • Legionella urine antigen tests only for Legionella pneumophila serovar 1, so as a whole it has about a 70% sensitivity
  • The histoplasma urine antigen is best in HIV+ patients with disseminated disease (95% sens); sensitivity is probably similar in non-HIV patients
    • It’s really not a great test for less severe pulmonary histo or chronic cavitary histo, especially in non-AIDS patients
    • It can be false positive with other endemic fungi (Blasto, coccidio, etc)
  • A brief note on IRIS
    • It represents ‘unmasking’ of an underlying, unrecognized infection as the immune system ‘reconstitutes’ itself while on ART
    • Many pathogens (TB, NTBMB, PCP, HBV, crypto) have been associated with IRIS, while others (ex: toxo) are rarely associated with it
    • Generally develops when pre-ART CD4 nadir <100, and generally develop within one week to a few months after initiation of ART (depending on the kind of infection, host characteristics, etc)


  1. Miller, Wally Jr and Shah, Rosita. Isolated Diffuse Ground-Glass Opacity in Thoracic CT: Causes and Clinical Presentations. AJR 2005.
  2. Kauffman, C. Histoplasmosis: a clinical and laboratory update. 2007.
  3. Murdoch et al. Immune Reconstitution Inflammatory Syndrome (IRIS): review of common infectious manifestations and treatment options. 2007.

8/17 PPMC Report: AL-Amyloidosis

Today we discussed the case of a patient who presented with macroglossia and was found to have AL amyloid and multiple myeloma.

We reviewed physical exam findings to confirm macroglossia — including scalloped tongue and a thorough differential for macroglossia including amyloid, hypothyroidism, acromegaly, space occupying lesions (abscess v. hemangioma v. lymphangioma v. cyst etc.) and exceedingly rarely GCA (1).

Types of Amyloidosis:

Type of Amyloid Description
AL Deposits monoclonal light chains – associated with plasma cell dyscrasias but can occur alone!
AA Deposits amyloid A – associated with chronic inflammation (ie RA, IBD, spondyloarthropaties), frequently presents with nephrotic syndrome
Dialysis Related Deposits beta 2 microglobulin, has an osteoarticular prediliction
Hereditary/Familial Deposits transthyretin, alpha cahin of fibrinogen A, apolipoprotein
Age Related (Senile) Deposits transthyretin – clasicially cardiac presentation
Organ Specific 2/2 locally produced proteins rather than circulating

The PPMC Report Team did a thorough screen for other end-organ involvement including renal (nephrotic range proteinuria), hepatosplenomegaly, CHF/restrictive cardiomypoathy, peripheral neuropathy, and carpal tunnel syndrome. We also discussed the increased risk for bleeding diathesis in AL amyloid patients due to both liver involvement AND direct factor X deficiency due to factor X binding to amyloid fibrils.

PPMC Team identified the disease specific signs/symptoms of multiple myeloma…the CRAB!


(hyperCalcemia >11, Renal dysfunction with Cr >2 , Anemia hgb <10, Bone disease >=1 lytic lesion)

We reviewed the different diagnostic criteria for plasma and lymphoplasmacytic cell dyscrasias (2):


We also reviewed the criteria for a patient with amyloidosis to qualify as having MM (3).  I may have misspoken during conference, so I just want to clarify here in this post: Patients with AL-amyloidosis who have >10% plasma cells on BMBx should be considered as having MM — Essentially pretend that smoldering myeloma doesn’t exist for AL-amyloid patients.

Our conference wrapped up with a discussion on cardiac amyloid and the medications in our arsenal to treat this disease: loop diuretics +/- spironolactone. Beta-blockers and CCB are contraindicated due to their negative inotropy and ACE-I are associated with significant hypotension (possibly unmasking some subclinical autonomic neuropathy). These patients should ALSO be considered for anticoagulation (with the help of a cardiologist) based on both rhythm and low atrial flow velocities.



  1. Helfrich DJ, et al. Giant cell arteritis of the tongue presenting as macroglossia. J Rheumatol. 1988 Jun;15(6): 1026-8.
  2. American College of Physicians. MKSAP 17: Medical Knowledge Self-Assessment Program. Philadelphia: American College of Physicians,  2016.
  3. Kourelis TV et al. Coexcitent multiple myeloma or increased bone marrow plasma cells define equally high risk populations in patients with immunoglobulin light chain amyloidosis. J Clin Oncol. 2013 Dec;31(34):4319-24.