Hepatic Portal Venous Gas (HPVG): a Less Ominous Sign than We Thought? A Case of HPVG associated with massive PE… #FOAMed, #FOAMcc

So a few years ago I had a patient in the ICU, post op for some abdominal surgery, and, using POCUS, I detected a hyper echoic area in the liver, in a wedge shape.  I scanned the patient and, lo and behold, there was a matching area of air-filled hepatic venous sinuses on CT scan. Well, my surgical colleague and I were very concerned and proceeded to inform the patient he would be needing exploratory surgery for what was likely ischémie bowel. He essentially – though in more polite words – told us we were idiots and that his belly felt fine and he didn’t think surgery would be needed at all.

His belly did feel fine. So were his labs. So we worried, but, given this whole thing about free will and consent, etc, couldn’t very well force him into what we felt was necessary surgery.

The next day he was fine. On POCUS, the area of air had shrunk. The next day, it was gone altogether.

We thanked him for his keen clinical acumen and for teaching us a good lesson.

However, we were a bit perplexed, because traditional teaching equated portal venous air with a severe bowel disorder, usually ischemic or inflammatory, with exceedingly high mortality. At least that is what we had been fed. We are both grads of 1999. Hmmm…

So over the next few years we saw a few of these cases, sometimes bad, sometimes not, and a review of the literature (see below)  showed an interesting evolution of the disease. Described in the 1950’s on plain films, hepatic air was a bad omen indeed, with mortality in the 75-90% range. In the CT era, the mortality started to “drop” to the 35-60% range. Now you can find quite a few reports of “surprisingly” good outcomes with conservative management. So this evolution doesn’t represent a change in severity so much as the technological capability to detect smaller and smaller amounts of air in the venous system – just increased sensitivity. And now, with POCUS – ultrasound is the most sensitive detector of air in a vascular tree – the associated mortality is likely to take another drop, not only because of our ability to detect very small amounts of air, but also because we are actually looking at the area, and also in a wider range of patient’ pathologies that those commonly associated with HPVG.


Clinical Case: HPVG and PE!

So a couple weeks ago I saw a patient in the ED who’d recently broken an ankle, had her foot put in a boot and managed conservatively and came back dyspneic and tachycardic. Here are a couple of clips:

As always, I start with the IVC:

Big & fixed.

Hepatic veins:

Biphasic flow.

Femoral veins:

So here the source of the problem is pretty clear, a large common femoral DVT.

She wasn’t very echogenic so I don’t have great clips of the heart but she had a dilated and hypocontractile RV with a McConnell’s sign (preserved apical contraction), small and hyper dynamic LV with septal flattening.

Now here is where it gets interesting, the portal vein:

You can clearly see bubbles traveling up the portal vein. Ominous, or not?

So clinically, her abdomen was normal, she had no abdominal symptomatology at all…


Pathophysiological musings:

So the severe RV obstruction resulted in significant venous congestion. Additionally, the decreased cardiac output – as manifested by a lactate of 4 and mild tachycardia/hypotension (110 HR, BP sys 90’s) was clear.

The etiology of HPVG in the literature isn’t clear – mucosal disruption, bacterial gas are all mentioned but as far as I could find, no definitive answer.

Is it possible that there is a “normal” inward leak of mucosal gas that is normally fully dissolved in the venous bloodstream, but that, in cases of low flow and/or venous congestion, the dissolution capacity (per unit time) decreases, and that gas comes out of solution?  Alternately, those who have increased intraluminal pressure (gastric distension, etc), the increased transmembrane gas driving pressure may overload an adequate blood flow…

This would explain the benign course of many patients, particularily those with gastric dilation.


Clinical course:

Based on hemodynamics, tachypnea and, to some degree, venous congestion, I decided to thrombolyse her using 1/2 dose lytics. Within a couple of hours her HR decreased to the 90’s and BP rose to 110 systolic.  Echographically, however, the IVC/RV findings remained similar, but the HPVG decreased. By the next day, HPVG was altogether gone, lactate had resolved and dyspnea was significantly better.


Take Home Message:

HPVG, although not quite as poor a prognostic sign as once thought, nonetheless warrants concern and investigation, even if the abdominal exam is entirely normal and without symptomatology, as correction of an underlying cause of “benign” HPVG (whether low-flow or bowel distension) would still need to be addressed.

In the meantime, I suspect that, reported or not, this has been noted by other POCUS enthusiasts, since we are now looking more frequently at this area, and are dealing with patients with low-flow states, congestion, bowel obstruction/ileus or more than one of these.

Hopefully some investigators will take a look at this phenomenon and delineate the pathophysiological mechanism!

Love to hear of your experience with this.




For those interested in POCUS, see here for a quick read primer on clinical applications of POCUS.


HPVG Review article 2009:



Bedside ultrasound case: Fibroids, Syncope and Dyspnea. #FOAMed, #FOAMus, #FOAMcc

So today, a 33F presented following syncope. She was mildly tachypneic wiyh a HR of 135 and BP of 130/80. I’m inserting the clip of my bedside ultrasound evaluation, as this takes place essentially simultaneously with my history-taking:

So this clip runs thru a few views, starting with an IVC long axis, showing a relatively plethoric IVC with minimal variation. This is not normal. Tells me to expect something abnormal downstream, unless someone has flooded the patient with IV fluids. The next view is the parasternal long, then short axis, showing an increased RV to LV ratio, and a small, hypercontractile LV, with septal flattening consistent with RV pressure overload, the “D” sign.  The apical 4 chamber follows with little else to add (difficult to measure TAPSE well in that segment).

So this is sure looking like pulmonary embolism, and I’m already toying with a half dose TPA, MOPETT-style, until the reveals that the cause of her starting oral contraceptives two months ago was to control heavy menses associated with large uterine fibroids… So I figure I’ll buy myself some decision time anyhow by ordering the CT angio – unless in pre-arrest, I don’t thrombolyse without formal confirmation – but I did start IV heparin on the echo findings. Here is the CT:

So this indeed confirms submissive embolism, particularly to the left PA.

Next?  I work in a community hospital, and although I’m totally comfortable thrombolysing PE, in this case, I was concerned about bleeding related to the fibroids, and I haven’t yet figured out a way to embolize bleeding vessels at the bedside, so I felt that the safest thing was to transfer her to a tertiary care center with a solid interventional radiology program. So off she went. I’ll update if anything funky was done like a catheter suction and I can get some clips.

So in terms of POCUS, I think this illustrates how speedily a diagnosis can be made, and although in this case the pre-test probability and index of suspicion was pretty high, it isn’t always!





For more POCUS tips, see here!

Bedside Ultrasound Case Debate Part 1: A Poll ! #FOAMed, #FOAMcc, #FOAMus

So I’m walking to the ED to reassess a COPD’er that was on BiPAP, and one of the ED docs sees me in passing and says – “I might have a case for you, she’s on her 3rd litre and still a bit hypotensive…I’ll let you know.”  So I re-route and decide to take a look right away, because I’m never fond of shock NYD.

So here is this woman in her 50’s, BP is 93/67, RR 22 and moderately dyspneic. She has been increasingly so for a few days without infectious symptoms. The X-ray is clear and her labs unremarkable aside from a lactate at 3.3 mmol/l.  She is moderately overweight but quite active. Non=smoker without any cardiorespiratory known illness and on no medications.

Here is what we see on ultrasound-enchanced physical examination:

So, what do you see?

In the first clip, we see a large, dilated IVC with little variation – despite the dyspnea, making it a more significant finding – according to the Effort-Variation Index (http://wp.me/p1avUV-9k).   This automatically implies there will be some pathology (unless iatrogenically very volume loaded) to be found downstream.

In the second clip, you have a hyperdynamic and underfilled LV and a dilated, poorly contractile RV.  In the absence of cardiopulmonary disease and in an active patient, this is highly suggestive of an acute process, namely pulmonary embolism.

On further questioning she had done a new yoga stretching class as a possible endothelial-damaging process.

So what did I do? Get a STAT angioscan:


What would you do next?


I’ll tell you what I did tomorrow, and hopefully have some good bloody arguments!


PS for awesome talks by amazing speakers, don’t forget to register for CCUS 2015!!! For more info: http://wp.me/p1avUV-aU and register at www.ccusinstitute.org!






Bedside Ultrasound: Quite a Case! #FOAMed, #FOAMcc

So here is an awesome clip from an ICU colleague of mine, Lorraine Law.  She was managing a post arrest (elderly woman who collapsed at home and was resuscitated but remaining in profound shock) case using bedside ultrasound and came across this pathology:

video courtesy of Lorraine Law & Shirish Shantidatt

what do you think?

scroll below for my thoughts…





So the clip starts with a subxiphoid 4 chamber view that clearly shows a massively dilated RV with a hyperdynamic and underfilled LV.

[For the hemodynamic novices, remember that the ventricles are kind of like roommates who share a pericardium. Especially in acute scenarios, if one gets overloaded, the other will have to give way, until the pressure equilibrates. If the process is exceedingly slow, they can do some renovations and stretch the pericardium, but this takes likely weeks. In this case, the elevated PAP overloads the RV and the RVDP > LVDP, resulting in decreased diastolic filling, which in turn drops the stroke volume/cardiac output/MAP.]

We can see that the RV TAPSE (tricuspid valve excursion towards apex) is really minimal, supporting an acute or acute on chronic process.

The clip then shows a long axis view of the IVC with echogenic material, most likely thrombus, with a to and fro motion, going in and out of the RA. Wow. You don’t see this very often.  The only thing preventing further travel is actually the fact that the cardiac output is so low due to massive embolism so that the flow can in fact barely carry the clots forward anymore at this point, similar to the sluggish IVC clip I put up a few months ago (http://wp.me/p1avUV-5t).

The most likely diagnosis is pulmonary embolism, and thrombolysis is indicated. Unfortunately despite my colleague’s timely diagnosis, the clot burden was likely too much, and despite thrombolysis, the patient passed away of intractable shock.  One can imagine that the TPA actually has to make it to the lungs, and with such a degree of obstruction, it is likely that very little actually got to the pulmonary vasculature…

Unfortunate case, but quite impressive images.

A crazy thought, using hindsight and with the luxury of knowing the fatal prognosis: intracardiac (RV) TPA bolus? Small spinal needle?  Anyone bold enough? Food for thought if (when) I see one like this…





Marco says:

Really quite impressive images. A couple of weeks ago I admitted a pretty young patient after a successful resuscitation due to massive pulmonary embolism. Immediately after ROSC in emergency department, he was transported to the cath-lab where TPA bolus was administered directly through a PA cathether. In ICU we continued the infusion. In less than 24 hours we obtained a relative hemodynamic stability and discontinued all the vasopressors, but the case remains unfortunate because despite therapeutic hypothermia the post-anoxic damage was so severe that led to cerebral death declaration two days later.


Thanks Marco, very interesting.  There is a recent study on catheter directed thrombolysis in PE reviewed at PulmCCM:(http://pulmccm.org/main/2014/randomized-controlled-trials/catheter-directed-thrombolysis-submassive-pe-better-heparin-rct/)

A physiological point about PE resuscitation is the relative inefficiency of CPR, as both venous return and LV filling is severely limited, so systemic perfusion is even worse than the usually poor output during chest compressions…

Thanks for reading!

Marco replies:

Thanks, Philippe!
The point about the possible inefficiency of CPR is crucial in my opinion. The patient I brought as example had a witnessed cardiac arrest (he called EMS when in respiratory distress) and CPR without interruption from the beginning, nevertheless he resulted in brain death declaration.
I remember very clearly a 43-year-old woman that 3 years ago had a massive PE in the OR shortly after a long lumbar vertebral stabilization. We admitted her to ICU after more than 80 minutes of CPR, a bolus of rTPA and with severe hemodynamic instability. RV was extremely dilated. When she eventually regained stability I had little hope about her neurological recovery, but surprisingly she was extubated the following day and last year she returned to our 12-months post-ICU follow-up showing perfect recovery.
I think that systemic and cerebral perfusion during “obstructive” cardiac arrests such as massive PE is very difficult to asses with current technology. A couple of times I was tempted to check it with trans cranial doppler, but usually there’s too much confusion during CPR.
When I was a resident I witnessed to a iatrogenic cardiac arrest in a patient with advanced monitoring that led to an interesting publication: http://www.researchgate.net/publication/10832333_Cerebral_perfusion_pressure_and_cerebral_tissue_oxygen_tension_in_a_patient_during_cardiopulmonary_resuscitation


Wow, very interesting cases.  What fortune to have been able to record that data, as obviously getting that in during CPR would be almost impossible.  TCD, at least after ROSC, could be contributory… Another option is using NIRS, which I’ll be working with this summer.

thanks again!


An Update on Pulmonary Embolism: NEJM’s PIETHO Study…what’s the verdict? #FOAMed, #FOAMcc

As has been discussed in a previous post (http://wp.me/p1avUV-7T), patients with sub-massive PE (hypoxic, tachycardic, some troponin rise, etc…but no hypotension) remain in a grey zone, which is, to me , a dubious situation at best – their mortality can be up to 15%, morbidity likely more.  Everyone agrees the low-risk patients don’t need thrombolysis, and everyone pretty much agrees that the patient in shock needs it.  There is data out there suggesting that some patients clearly benefit from thrombolysis despite not being in shock, in good part relating to avoiding chronic pulmonary hypertension and its consequences.

The issue for many clinicians is that they have a “stable” patient in front of them, and they are considering giving them a drug that can potentially give them a bleed in the head and leave them dead or crippled. Many shy away from this. Part of this is cultural, because the same docs probably wouldn’t hesitate giving the drug to a lateral or posterior MI, which is not likely to kill you, or even leave you a cardiac cripple (just to be clear, I’m not advocating against thrombolysis in these cases, just trying to find a parallel), but since the AHA guidelines say to do it and everyone else does it, there’s no trepidation. It is the standard of care.  For most of us acute care clinicians who do not do outpatient medicine, if the patient survives and gets discharged home, chalk one up in the win column. But, as has become clear in recent years with the post-critical illness syndromes, morbidity can be just as important as mortality, especially in the younger patients. Kline et al (Chest, 2009) showed how almost 50% of “submassive PE” patients treated with anticoagulation alone had dyspnea or exercise intolerance at 6 months. They only had a 15% improvement in their pulmonary artery pressures (mean 45 mmhg).

What are the real risks? Pooling the data together gives a value around 2% with a spread between 0.8% and 8%, more or less. This represents each patient’s inherent risk of bleeding, as well as some of the inconsistencies with post-thrombolysis anticoagulation (safest to aim for 1.5-2 x PTT baseline in the first 48h).

The MOPETT trial which, as a #FOAMite you have certainly come across, showed that a half-dose of TPA was highly effective, and they felt it might be possible to go lower. The physiological beauty in that is that, unlike other sites we thrombolyse with full dose TPA, the lungs get 100% of the TPA (coronary artery gets maybe 5%, brain gets 15%).  Mind you, of course, the culprit clot/artery obviously doesn’t get 100%, but much, much more (if we figure that you need about 50% vascular area occlusion to cause RV dysfunction) TPA per “clot” than other pathologies. One can argue that anatomically, there is a greater clot burden than coronary or arterial thrombolysis, which may offset this somewhat. However, the date was quite clear in this trial that the therapy was effective, and the bleeding was none.

Ok, so let’s get to the PIETHO. 1000 patients, TPA+heparin vs heparin alone in normotensive but intermediate risk patients. So, first question is how was that risk defined?  Patients needed to have echocardiographic/CT signs of RV dysfunction AND a positive troponin. Interestingly enough, onset of symptoms was up to 15 days before randomization…not exactly early treatment, and unfortunately there is no information about the actual time to thrombolysis or subgrouping.  The results were as one could imagine. The combined endpoint of death or hemodynamic decompensation was 2.6% in the thrombolytic group vs 5.6% in the anticoagulation.  I’m not a fan of combined endpoints. Hemorrhagic stroke was 2.0% vs 0.2%. Their conclusion? Exercise caution. Hmmm…not much of a step forward. Basically tells us what we know. It helps the hemodynamics, but you can bleed. They do re-affirm that bleeding is more likely in the over-75.


What do we REALLY need to figure out? 

1. echographic risk stratification – at least into moderate and severe RV dysfunction.

2. longer term outcomes (hopefully PIETHO has a follow-up study in the pipeline, since they had good numbers).

3. a point-of-care study – time is of the essence, and may have an impact on dosage. IMHO thrombolysis should be done within a few hours of presentation at most.

4. further dosage data – 1/2? 1/3? 1/4? small boluses q1h until RV function improves?

I wish I could do it, but community hospitals don’t have the ideal setup, nor do I have a research team that can handle something of this scale. But surely someone can!


Bottom line?

It won’t change my practice. I will continue to offer thrombolysis in select cases, especially the younger patients, who obviously have a lower risk of bleeding, and stand to benefit the most, as pulmonary hypertension  can be crippling. I know I’d take the risk of bleeding when I see 50% dyspnea/exercise intolerance two years down the road…

Finally, bedside ultrasound to anyone with dyspnea/hypoxia should be a standard of care for every acute care physician. No ifs, ands or buts, no exception. Waiting for a CT angio or formal (read daytime hours) echocardiogram is, to me, unacceptable. If you, a friend or family member were in that ER bed, would you trust a physical examination and a CXR to rule out the need for an immediate intervention? I wouldn’t, not my own, and not even Dr. Bates’, Dr. DeGowin’s or Dr. Sapira’s, or all three combined.





Kline JA, Steuerwald MT, Marchick MR, Hernandez-Nino J, Rose GA. Prospective evaluation of right ventricular function and functional status 6 months after acute submassive pulmonary embolism: frequency of persistent or subsequent elevation in estimated pulmonary artery pressure. Chest 2009;136:1202e1210.

Guy Meyer, M.D., Eric Vicaut, M.D., Thierry Danays, M.D., Giancarlo Agnelli, M.D., Cecilia Becattini, M.D., Jan Beyer-Westendorf, M.D., Erich Bluhmki, M.D., Ph.D., Helene Bouvaist, M.D., Benjamin Brenner, M.D., Francis Couturaud, M.D., Ph.D., Claudia Dellas, M.D., Klaus Empen, M.D., Ana Franca, M.D., Nazzareno Galiè, M.D., Annette Geibel, M.D., Samuel Z. Goldhaber, M.D., David Jimenez, M.D., Ph.D., Matija Kozak, M.D., Christian Kupatt, M.D., Nils Kucher, M.D., Irene M. Lang, M.D., Mareike Lankeit, M.D., Nicolas Meneveau, M.D., Ph.D., Gerard Pacouret, M.D., Massimiliano Palazzini, M.D., Antoniu Petris, M.D., Ph.D., Piotr Pruszczyk, M.D., Matteo Rugolotto, M.D., Aldo Salvi, M.D., Sebastian Schellong, M.D., Mustapha Sebbane, M.D., Bozena Sobkowicz, M.D., Branislav S. Stefanovic, M.D., Ph.D., Holger Thiele, M.D., Adam Torbicki, M.D., Franck Verschuren, M.D., Ph.D., and Stavros V. Konstantinides, M.D., for the PEITHO Investigators*, Fibrinolysis for Patients with Intermediate- Risk Pulmonary Embolism, N Engl J Med 2014;370:1402-11.

Mohsen Sharifi, MDa,b,*, Curt Bay, PhDb, Laura Skrocki, DOa, Farnoosh Rahimi, MDa, and Mahshid Mehdipour, DMDa,b, “MOPETT” Investigators, Moderate Pulmonary Embolism Treated With Thrombolysis (from the “MOPETT” Trial), Am J Cardiol 2012

“Doc, I can breathe!” – Thrombolysis in PE…a case discussion. #FOAMed, #FOAMcc

So I was on call last weekend and got a call from one of the internists on the ward about a potential admission who may need dialysis.   She was a woman in her 60’s, diabetic, hypertensive with minimal baseline renal dysfunction, who had been admitted with a hepatic abscess due to biliary obstruction. This had been stented and a pigtail catheter had been inserted to drain the abscess.  However, over the last few days, her creatinine had risen to about 500 and she was becoming oliguric.  Her O2 requirements had also increased and she was now on 15 liters by nasal prongs. This had been ascribed to pleural effusion and possible pneumonia.

When I saw this lady, she was visibly dyspneic at 30 with a heart rate 115-120 and a systolic BP of about 105-110, saturating 90% on 50% face mask.

So on physical examination, she had a soft abdomen (the first thing I feel just before I put probe to skin), her skin was cool, and the CUSE revealed a large (>20mm) IVC with no respiratory variation (despite the effort).  I unfortunately forgot to hit the record clip button…and the parasternal long axis and apical 4 chamber are here:

Lung views showed “A” profiles except for the right base which had a small effusion and some consolidation/atelectasis and some B lines, but not very extensive.

So further assessment revealed she was not a smoker, previously quite active and easily able to go up and down several flights of stairs.  She had noted dyspnea about 3 days ago, without chest pain. There were no leg symptoms, and she had been on LMWH for dot prophylaxis.  The CXR was not very impressive – in a sense that there was not enough parenchymal disease to explain pulmonary hypertension.

This is PE until proven otherwise, and I would have been comfortable without further confirmation, but with the presence of some lung disease and an intrahepatic catheter, I preferred to have 100% confirmation before initiating thrombolysis.

After CT angiogram confirming bilateral and extensive embolism, I had a thorough discussion with her and her family and they all agreed to go ahead with TPA.  She was quite concerned with cardiorespiratory limitation, given that she was quite active. She was comfortable with a quoted risk of intracerebral bleeding below 2%. I used the MOPETT half-dose of 50mg.

Overnight, her HR slowed to about 100, and sats increased to 93-94%.

When I rounded on her in the morning, she said “Doc, I can breathe!” with a big grin. Her HR was 95-100, she was not on 3 litters by NPs, BP 115-120 systolic, and CUSE showed:

So we can see that even though the RV is still quite impaired, it has decreased in size and the LV is now filling better. This was about 12-13h post thrombolysis. She was able to sit up without dyspnea and mobilize to the chair. Her IVC, although it remained around 18-19 mm, had clear respiratory variation.

So…success? Who really knows. It is concievable that, with heparin alone, she might have improved similarly. It is possible. I’m not putting this up to formally support the concept of thrombolysis in “submassive” PE but more to contribute to the #FOAMed discussion regarding the “grey zone” of thrombolysis, since she was technically not in shock (eg SBP>90, lactate normal), but the degree of impairment of the RV to me and the clinical picture, 3 days post, was concerning enough to warrant thrombolysis, but importantly to stress the following:

Point 1: the importance of bedside ultrasound, especially in acute cases.  Without it, over a weekend, and with a patient in renal failure, how quickly would I have ordered a CT angio?  Not without some hesitation…

I won’t review the MOPETT trial, these guys did a much better job than I could hope to, so definitely listen to this if this topic is of any interest to you (and it should!!!):



Great case debates in the RAGE podcast.

Keep in mind that morbidity, not mortality, is the main thing to focus on in sub-massive embolism and the MOPETT – even though I don’t really like the term, its quite vague – benefit in embolism with shock is quite clear.

Point 2: Equally interesting to me was the fact that the renal failure improved. In fact, overnight following thrombolysis, she had a urine output (without diuretic) over a litre, and over the next few days her creatinine normalized and renal replacement therapy was not needed.  Interesting, since she even got a good blast of toxic dye with the CT.  Some will feel that it is the improvement in CO that improved renal function, and this may be partly true, but in view of the lack of “systemic shock,” I think that venous decompression resolved the congestive renal failure, which I think was the main cause of her ARF. I posted about this topic a few months ago, so for more on this see:


so thanks for reading and love to hear anyone’s opinion!





Great question!  There is a whole grey area in “PEA” and management is unclear. I don’t think there is a single answer to that, but physiologically and without further information about RV/LV, I would say your patient needs vasopressor/inotrope support, so I would probably give a small bolus of epi (maybe 100ug) and start an infusion. If I see little reaction (eg HR/BP doesn’t pick up in 30 seconds, I would probably give a short cycle of CPR to get the epi back to the heart.   Of course, hopefully there is a reversible cause (MI/PE), that can be addressed.