Podcast 111 – Fluids in Sepsis, A New Paradigm – Paul Marik


Dr. Paul Marik is a renowned intensivist and a confirmed critical care skeptic. He has broken down many myths such as the use of CVP for volume assessment. I recorded a lecture he gave to the Sepsis Collaborative I co-chair. It is an amazing lecture.
I agree with some of it and have quibbles with other parts, but it is a must-listen. In one week, I will publish a wee with the areas I see differently; in the interim think about your own viewpoint so that we can discuss it all in the comments.

Dr. Marik’s Previous Visits to EMCrit

Dr. Marik’s Updated Hemodynamic Management Flowsheet

Marik Fluid Flowsheet

Dr. Marik’s Slideset

Now on to the Podcast…


  1. Alex K says

    Awesome lecture, enjoyed immensely.
    Pretty sure my hospitals don’t have NICOM. Have been following the ultrasound podcast closely regarding their thoughts on measuring cardiac output with the PLR, most recently exploring the measurement of carotid pulse wave dynamics. Wondering if this might be a good compromise?

  2. Roger Helmers says

    The podcast on fluid therapy in sepsis was the most mind-blowing lecture on critical care I have heard in my life. Some of the things I have been doing for years (e.g. avoid the “Abnormal Saline”) but some was just jaw dropping new and astounding. I will have to listen to it at least half a dozen times.

  3. says

    Thanks for sharing Scott!
    I’ve just had a relapse of my Marikophilia…
    One of my favourite lines in his textbook reads: “November 8th 2001, one of the darkest days in the history of critical care…” (the day both the Rivers and the Van den Berghe were published)

    • Matthew Mitchell says

      Editors Note I have removed the comment posted by Matthew Mitchell that resided here.
      This is now the third post I have had to remove in the history of EMCrit. WE DO NOT ALLOW AD HOMINEM ATTACKS on the EMCrit site. We also do not allow postings by people without a name, an affiliation, and a COI statement in the comment itself. I let this slide when people are obviously commenting in beneficial ways. Mr. Mitchell, if you want to post a revised comment that complies with the above, I welcome it.

  4. says

    Exellent lecture,
    there are some game changers in it. Especially I liked the mention on beta blockers in septic shock. In my personal experience a tachycardic heart rate is a very predictable marker of badness. The use of dobutamin even if you did a reasonable fluid management remembers me of a ancient roman galley. You just command the drummer to increase his beat to gain speed.
    Another very good model is the MARIK curve.

  5. Len Ulan says

    Scott, Interesting talk by Dr. Marik. I look forward to the author of “Emergency Medicine Decision Making” applying his critical eye to the evidence used in this lecture. I suspect he will find many of these studies are poorly designed and underpowered.

  6. Paul says

    He mentions over and over again the studies that consistently show that patients who had the highest fluid balance died compared to patients with lower balance that lived. Maybe I’m missing something and being overly simplistic here, but couldn’t you say that maybe the patients who were sicker, had poorer MAPs/indices of perfusion etc… were the ones to likely receive more fluid? Sure, the patients whose MAPs were improved with 4 liters of fluid are probably in better shape than ones who don’t respond to the same amount.

    • says

      yes, one could say that, and I will in 1 week. Further, let’s not equate fluid balance with fluid administration–something to think about for the wee.

  7. Mike Jasumback says

    All of which makes me want to volume load my next hypertensive decompensated CHF pt! Another brilliant lecture
    Thank You Scott and Dr. Marik

  8. George Hutchison says

    So is Rivers EGDT another ‘steroids in spinal trauma and TPA in stroke’ all over again or is this guy another crazy man who turns out to be correct a la Mad Dog Madison regarding fluid resuscitation in trauma ? Realizing that EGDT is currently recognized as the standard of care with medico-legal and, more importantly, patient care ramifications the answers need to come soon. CHI, for example, is championing a national conquering sepsis campaign based on EGDT.
    One glaring concern in Dr. Marik’s algorithm is his lack of a clinical marker of perfusion. Using MAP alone is not adequate as has been so eloquently pointed out in past lectures regarding resuscitation.
    Lastly, I think question 2 in the post test is ambiguous. Norepi reduces physiologic venous volume (reduction in venous pooling by venoconstriction) leading to improved preload. Effective fluid volume is therefore increased.

    • says

      The clinical marker of perfusion that we use is DO2I, and VO2I as the marker of tissue oxygen usage. Why use surrogates like ScvO2 when you can use the real thing? We also look at CO, CI, SV, SVI, inotropy, preload, afterload, SVR, PKR, and a half dozen others. It really just makes the treatment of sepsis like painting by numbers! Do you want to use an “off the peg” protocol or go with “made to measure”?
      Incidentally, if anybody would like a copy of our algorithm (BUSH protocol) then just flick us an email.

  9. says

    Wow, as an aspiring EMCCM doc I thought I was just about to grasp the slightest bit of an understanding of septic management and then this! My jaw was literally dropped. Can’t wait for your Wee-sponse.

  10. Craig H says

    Outside of an amazing talk, basically, it supports the idea that generalization of therapy/treatment for any dynamic state, and more so fragile dynamic states, require individualized care and repeated bedside assessment… seems pretty much summed up by Leith, in 1832. Adds fuel to the fire that “standard” therapy for QI/QA measures may get you positive metrics but poor outcomes.

    Either way, great post as always and wonderful talk.

  11. roslan says

    thank you for the excellent stuff…it sort of confirm that septic shock has more than meet the eyes:

    few observation:

    1. dr marik algorithm is almost similar to BUSH protocol introduced in bathurst hospital australia by Smith…he used USCOM as a noninvasive monitoring…. 3 years data in his hospital show mortality rate lower than other tertiary hospital in Australia. By using the protocol he managed to stabilized hemodynamic including cardiac index, vasodilation, inotropy index within 60 minutes compared to normal EGDT

    2. he also introduced inotropy index to measure contractility (Smith Madison Inotropy Index)SMII. What do you think?
    Smith BE, Madigan VM. Non-invasive method for rapid bedside estimation of
    inotropy: theory and preliminary clinical validation. (Br J Anaesth. 2013
    Oct;111(4):580-8. doi: 10.1093/bja/aet118. Epub 2013 May 3. PubMed PMID:

    • says

      I agree.

      I had the pleasure of having Prof Brendan Smith come give a talk to my emergency department. We also use the USCOM. Both he and Dr Marik share the same message with regards to fluid therapy; you really don’t need that much. I find the article where giving noradrenaline early in septic shock patients increase their preload and cardiac output most fascinating.

      With regards to the SMII, we got the software upgrade to the USCOM recently and I have yet to try it out in shocked patients. Should be interesting.

  12. Rob Phillips says

    Fluid is good if you need it and bad if you don’t. Fluid responsiveness is SV increase associated with volume expansion, so you have to measure SV reserve and that requires an accurate SV measure (or more usefully SVI). Importantly in sepsis where 50-100% of subjects have myocardial depression its important to avoid fluid loading myocardial depressed normovolemics. Which brings us back to SVI. USCOM is the quickest and most accurate method for fluid guidance and works in AF, on and off ventilation, with vasoactives and in sepsis with active neurohumeral activity and changing SVI and SVRI. (Thiel, Crit Care 2009) Brendan Smith’s work proves this. Using BUSH protocol USCOM goals of SVI, SVRI, CI, SMII and DO2I mortality in fluid resistant septic shock was reduced to 11% in children (Deep, Int Care Med 2013). USCOM transcutaneous Doppler with physiological haemodynamic objectives works, and makes the administration of fluids, inotrope and vasoactives rational.

  13. Dave Barounis says

    Hey Scott,

    I think the lecture is very interesting and I really appreciate you sharing it. I think some of this is already being practiced, but there are a few things I am fairly certain you are going to touch on in the wee but I thought I would drop in here:
    Dr. Mark discusses fluid balance and that a positive fluid balance has been identified as a marker of worse outcome, which he is attributing to over zealous fluid resuscitation. However, patients who have a positive fluid balance tend to be sicker (i.e; they got more fluid because of severity of illness), and their fluid balance may be positive because of renal failure (UOP sucks) for which there is good data to suggest is an independent predictor of mortality. The problem, as I am sure most of your listeners know, with retrosepctively adjusting for the severity of illness is patient A could have a lactate of 4, which has cleared from 8 and is starting to make urine (and his Cr is improving to X) while patient B could have a lactate of 4, increased from 2 and has reduced urine output with a Cr increasing to X. Non-randomized, retrospective studies can never account for these changes when calculating severity of illness scores since they often use static numbers.

    The next point is calling normal saline devil’s fluid. Almost all of these studies were done outside of the ED and in the ICU (usually after a significant portion of fluid resuscitation had already occurred.)

    -The SAFE study compared NS and 4% albumin with no difference in any of the pre-specified primary outcomes ( he does mention using 20% albumin instead, which to my knowledge no RCT’s exist documenting this, although I buy it makes sense).
    -The CRISTAL study actually said that Colloids had improved 90 day outcome but the main colloid used was HES (which most people had written off from the CHEST study).
    -The next paper which he mentions was using chloride rich versus chloride restrictive fluids, which did show an increase in RRT during their ICU stay with an increase in serum creatinine requiring dialysis. However, mortality, RRT after discharge from the ICU, ICU and hospital LOS were the same.

    As many intensivists know the decision to put someone on RRT is not always clear cut. Do I still use LR as my resuscitation fluid of choice, usually, but mostly so I am not trying to fix their ensuing acid-base disorder which may or may not have clinical relevance. I don’t , however, think NS is killing patients by the 100’s. A good read is the editorial by derek angus and Christoipher Seymnour on fluid resuscitation in JAMA http://jama.jamanetwork.com.proxy.cc.uic.edu/article.aspx?articleid=1752242

    Probably in the end, giving “just enough IVFs”, pressors, and picking whatever crystalloid seems to make sense at your institution is a very reasonable approach. I agree MOSTLY with Dr. Marik, but I definitely don’t think its gospel. Thanks again.

      • says

        Here is the citation for the editorial:

        SAFE compared two unbalanced solutions NS vs. albumin both are NS, only difference is the albumin.

        I’ve seen nothing compelling that any colloid except albumin has no role in resuscitation at this point. I would love to see a balanced albumin product. There are now at least 3 studies demonstrating unbalanced is associated with worse outcome, but these are merely associations and hopefully ANZICS can do a real trial on this similar to the SAFE methdology.

  14. Rob Phillips says

    You’re right Dave, but to give “just enough IVFs” and pressors requires an accurate measure of SVI and requires the monitoring of the SVI response. USCOM has a sensitivity of 5% to changes with inos and vasos compared to flow probes. (Phillips, Crit Care Res Prac, 2012) Other technologies can’t provide that sensitivity, and so are limited for monitoring and optimising therapy.

    • Francesco Savelli says

      I agree with Rob, I’ve used the USCOM in several case of severe sepsis and I find it’s really very useful to test the response to fluid challenges. Usually I don’t use the passive leg raising test, if I think to be close to the top of the Frank Starling curve I prefer a mini fluid challenge with 250 ml of saline in ten minutes, I find it more pratical. Anyway the monitor has been validated for passive leg raising tests (Thiel Critical Care 2009; Wang HL Zhongguo Wei Zhong Bing Ji Jiu Yi Xue. 2011 Mar;23(3):146-149) and I don’t understand why Marik has not included the USCOM, together with esophageal doppler, NICOM and calibrated pulse contour devices, in his slide on this topic. Furthermore since I’ve upgraded the monitor to the inotropy index, togheter with the sonographic evaluation of the caval index, I’ve realised that not few patients probably don’t need fluids since the beginning of their admission to the emergency department and should be treated earlier with inotropic drugs

  15. says

    Mr. Phillips,

    I believe you have a few appropriate disclosures that may put your comments in the proper frame for the readers. It doesn’t mean anything you are saying is less than true, simply that you should make folks aware. I will leave that to you or I will have to remove these comments.

    Mr. Savelli–I know you have spoken on USCOM, but do you have any conflicts of interest when discussing the device?

    Thank you both for commenting

    • Francesco Savelli says

      Dear Dr Scott,
      you are right when you write that I’ve spoken about the USCOM and I’m going to talk again about this device in Italy because I’m trying to elicit interest between my collegues Italian emergency physicians about noninvasive methods for cardiac output monitoring. But, as I’ve talked about the USCOM, I’ve talked also about the Nexifin and the NICOM, and I can assure that I haven’t any conflict of interest with the USCOM company (but may be I’ve some conflict with my wife for the time I’m spending in this clinical-cultural interest!). I had the lucky to try all this three devices and, in my small experience, I can say that the ideal device doesn’t exist. Nexfin is very easy and it is the best to check postural changes in cardiac output, but often doesn’t work in the critical patient or it’s unreliable, being a non calibrated system. The NICOM seems the best but it gives few information about contractility (the dX/dT seems to reflect the dP/dT but to my knowledge it hasn’t been clinically validated) and also this device, like the others, recently had a negative evaluation (Br J Anaesth. 2013 Aug 28. Bioreactance is not reliable for estimating cardiac output and the effects of passive leg raising in critically ill patients. Kupersztych-Hagege E, Teboul JL, Artigas A, Talbot A, Sabatier C, Richard C, Monnet X). The USCOM is not suitable for the evaluation of postural changes in CO and in old people it can be difficult to get a good signal for the aortic valve from the suprasternal notch, but its information about inotropy are unique. The cardiac power, which is given also by the other devices, is similar as concept to the inotrpy index, but it isn’t the same because it takes into account only the absolute value of cardiac output (multiplied by the arterial BP), not the kinetic energy developed by the heart, and until now it has been validated only for its prognostic value. However, since I’ve had much more time (and therefore satisfaction) to experience the USCOM, compared to the other instruments, this has probably influenced my judgment and I apologize for that. But I believe that if we want to carry on a management of sepsis less invasive as possible we can’t base our decisions about fluid resuscitation only on caval index, but we have to consider noninvasive CO monitoring, regardlessof the instrument that we have chosen.

      • says

        Comments such as these from someone not taking speaking fees from a company (that is what I am interpreting from your email) are very helpful. I unfortunately have not had an opportunity to trial the USCOM as of yet. From looking at the videos, this is a device that requires an operator to acquire a signal for each set of readings, correct? That would seem the only disadvantage, the same disadvantage I see with point-of-care ultrasound. The ideal would be something we could put on or in and just get the readings without a clinician having to take the time to get each set.

        • Francesco Savelli says

          Thanks a lot for the consideration given to my modest opinion ,
          you have interpreted correctly my words, I didn’t take any money or any other kind of fees to speak about the USCOM. You are right in comparing the USCOM to the “point of care ultrasound” and in my clinical practise I always try to do both, because I think that the information given by this two methods complete each other. After “fast ultrasound”, I like what I call the “fast doppler” . If the patient is easy to examine it isn’t a problem to make repeated measurements to have a trend, it takes really few time. The real problem are patients with dyspnea where the suprasternal window is less suitable, but you can try the pulmonary valve trough the parasternal approach.
          I hope you can try soon the device and give us your valuable opinion, your blog is really great!

    • Rob Phillips says

      Hi Scott. Thanks for the opportunity to frame my comments in the context of my association with the development of USCOM. I have had 20yrs of clinical, research and education experience with echocardiography and have completed 2 theses on ultrasound evaluation of cardiovascular performance. The USCOM device was developed specifically to improve on the weaknesses in other methods and so the science to me is very important. Further I believe that sounds science leads to best practice, so my contributions were based on science and not opinion and I welcome the contribution of agreement and disagreement, and any evidence of unscientific bias in my perspectives. The biases we all have are a product of our learning, and by intelligent interaction this increases; this is the academic process and Im pleased to participate in that.

      • says

        Mr. Phillips,

        I’m sorry if I didn’t make myself clear. If you would like to comment on USCOM here, and I hope you continue to do so. Please acknowledge that you are on the board of directors and the executive chairman for the company that manufactures the device when making such comments. This is just good practice in the FOAM community. I would have thought this request would have been understood from my prior comment.

        • Rob Phillips says

          Thanks Scott. I am on the Board and executive Chairman of Uscom, facts which don’t change cardiovascular physiology or the validity or otherwise of opinions discussed above. I look forward to this valuable and overdue conversation continuing and thanks for the opportunity to participate.

  16. graeme Pickford says

    Loved the podcast. It is a big conceptual leap to go from current practice to “as much fluid as you need and not one drop more”, and most ED docs cant measure this easilly anyway. BUT, as Paul Marik suggests, it’s not a big conceptual leap to see that 11 or 12 litres x NS is bad–and yet my guess is that I’m not the only one who sees this all too often. This is an easy middle ground. Care with fluids rather than liberal fluids is a good starting point?

    • Ben Dowdy says

      I was actually just going to ask a similar question from the prehospital perspective. With a high clinical suspicion of sepsis based on history and hypoperfusion (heck, toss in a portable lactate analyzer for good measure), at what point should paramedics begin to add in pressors/inotropes? Does hypotension refractory to a 30mL/kg bolus suffice?

      • says

        We would say that if the patient has not responded to a 20ml/kg fluid challenge by increasing systolic blood pressure by 10 mmHg, then you probably have a flat Starling curve and should be thinking of alternatives to fluid resuscitation. One exception to this would be in those patients where there is clearly an obvious element dehydration which may have developed over the preceding day or two. In this context, the hyperglycaemic patient who has had excessive polyuria, or the patient found collapsed on the floor after an extended interval of downtime, springs to mind. In these patients you will obviously need to make good the backlog of fluid as well as giving a little more as a volume challenge.

      • roslan says

        If my interpretation are correct, if you use dr marik algorhythm u should start norepi after 30ml/kg fluid bolus….this will shorten the time to achieve hemodynamic stability compared to using EGDT / SSC2012 (3hr)…..

        Which goes back to the question if we aggressively treat our septic shock patient so that we can shortened time to achieve hemodynamic stability, then we can reduce mortality even more.Egdt manage to reduce it by three hour which give ARR of 17% what if the next frontier in sepsis management is to reduce it even further, let’s say 60 min…..I think this is what dr marik and BuSh protocol aims…..

        BTW I’m just a normal EP from malaysia without any ties with USCOM..I’m just happen to visit the booth during one of the critical care conference and try it on site….my first try I can get the wave in less than 10 sec…..definitely easier to do than your US estimation of cardiac output …TQ

  17. Rob Phillips says

    Thanks EMCRIT. My contributions to this discussion are framed by my direct experience with the development of the USCOM device. I should add I have spent 20yrs in echocardiography teaching, research and clinical practice and have completed two theses on the topic of ultrasonic evaluation of cardiovascular function. So while I am conflicted by association I am comfortable regarding the quality of the commentary and science and recognise the priority of sound science in the development of best practice, and welcome intelligent engagement with contrary or complimentary perspectives.

  18. Rob Phillips says

    As an aside and in the interests of full disclosure Uscom does not pay speakers or researchers because of the potential for conflicts of interests. Paid or biased opinion is compromising for us all. What is written and said on USCOM reflects real science and real beliefs. We are an old fashioned science company trying to contribute to clinical care and education.

  19. says

    Once again, Paul Marik is to be congratulated on bringing good common sense mixed with academic rigour to a treatment modality which we think should be consigned to the history books, i.e. throwing litres and litres of saline at septic patients.
    Just to deal with earlier comments, it is not simply a case that sicker patients require more fluid and therefore those patients with higher fluid requirements have higher mortality. In the BUSH protocol study to be published early in the New Year, we found that in propensity matched patients, mortality increased sharply with a fluid balance in excess of 4L in the first 24 hours or 7L in the first 48 hours. Those patients treated by the BUSH protocol had a mortality of just over 6% compared with 38% with those treated by what might be termed conventional EGDT methods. The non-BUSH group also had a typical positive fluid balance in excess of 13L at 48 hours as opposed to under 7L for the BUSH group. We agree with other commentators that having an accurate means of measuring the haemodynamics is essential. In this study we used an USCOM, and frankly, having used just about every other haemodynamic monitor out there, we will not be going back to using anything else.
    Now before continuing, and mindful of emcrit’s comments above, we must state that we have no financial interest whatsoever in USCOM; we’re not consultants for USCOM, nor paid by USCOM, we do not have stock or shares in USCOM, and as far as we’re aware, no member of our entire extended families has ever worked for USCOM! We do however love their device. All we ask in a clinical monitor is accuracy. The fact that it is not a continuous monitor doesn’t bother us at all, because our treatment is not continuous. We assess patients, we tweak the treatment, and then we review the patient after a period of time. We remeasure the haemodynamics at that point, and the cycle starts again if needs be. There is no advantage to having continuous monitoring with intermittent treatment plans, especially if that continuous monitor needs to be recalibrated before treatment decisions are made based on the results that it provides .
    One of the most fundamental points about fluid therapy is being overlooked in septic patients. We found that 89% of our patients with septic shock had significantly depressed inotropy as measured by Smith Madigan Inotropy Index (SMII). This means that they are on a flat Starling curve, and as such they cannot respond to fluid. To keep throwing fluid at the hypotension is just as ridiculous as throwing fluid at cardiogenic shock. The measurement of ejection fraction in widely vasodilated patients is also meaningless, as any old heart can eject a good fraction of its LVEDV when there is no afterload! In our experience, the best predictor of volume responsiveness is SMII and starting preload level prior to fluid challenge.
    We also have to question the whole concept of “massive capillary leak syndrome”. Whilst this certainly occurs in late septic shock we are far from convinced it is present in the early stages. If it were, then patients would present rather like anaphylactic shock patients with swollen hands, eyes, lips, tongue, larynx and airways. The haematocrit’s would all be sky high and the haemoglobin levels over 20g/dL. They are not. We create the extensive systemic and pulmonary oedema by adding salt water drowning to the pre-existing problems of an impaired left ventricle in a very sick patient. That is why so many patients develop haemoglobin levels below 10 g/dL. The sooner we get away from giving vast quantities of fluid based on meaningless pressure measurements such as CVP, PAOP, or vena caval distensibility (which is just a difficult way to estimate CVP!), or on surrogates of cardiac output such as ScvO2 or lactate levels, the sooner we can start treating these patients scientifically, and reducing the world mortality to the kind of levels that the BUSH protocol and its derivatives are achieving.
    Paul Marik has done the Emergency Medicine and Intensive Care worlds a great favour by challenging the dogma. Far too many patients have already died from this flawed creed; it’s about time we let the light of science shine in.

  20. dave barounis says

    While I am very intrigued by the use of USCOM, and admittedly have not used the device, I am suspicious of any diagnostic test that supposedly reduces mortality by 98%. Since the therapeutic interventions (IVF judiciously, ABX early and broad, and pressors +/- steroids) haven’t changed, is the argument that optimization of the aforementioned treatment modalities has the possibility to reduce mortality by 94%?

    My skepticism comes from the teaching that no diagnostic tests without appropriate coupling to therapeutic interventions has improved mortality (i.e; cxr, CT, MRI). This is not to say I don’t think hemodynamic optimization is important and can improve mortality, just that a 94% change seems like a bit of a stretch.

    Also is this paper going to be published in non-abstract format? Thanks for the discussion.

    • Brendan Smith & Veronica Madigan says

      You’re dead right to be suspicious of any diagnostic test that supposedly reduces mortality by 98%!

      I would just like to put the results from the BUSH study in perspective. Those patients treated according to the results of the haemodynamic measurements performed immediately after presentation, and whose continued resuscitation was guided by haemodynamic results, which we shall call the BUSH group, had an overall mortality of just over 6% as opposed to those treated by conventional methods where the mortality was 38%. That isn’t a 98% reduction, but it is still pretty impressive we feel! Where the confusion comes from is that our mortality from septic shock back in 2006 at the start of our study was almost 60%, whereas in the treatment group it is now only around 10% of that. Back then, we really weren’t instituting any appropriate therapeutic measures, not even early antibiotics, let alone what might be called the full bundle of therapy. However, clearly there is still a great deal of improvement that can be made on the normal bundle of care.

      Just to return to the theme of Paul Marik’s podcast however, one very important aspect regarding fluid that we would like to comment on is simply the question of “where is the fluid that you give now going to be in one hour’s time?” If you pour in 6L of crystalloid over a couple of hours, then at most, only 2L of that fluid will still be intravascular in an hour or two from now. Where will the rest be? If we’re lucky, it may be in the bladder or the catheter bag. More usually it will be in the tissues as systemic oedema or even more alarmingly, as pulmonary oedema. Meantime of course, because the circulating volume has dropped considerably, the patient will now need further fluid to restore cardiac output and blood pressure. The cycle will then repeat and we’ll all just end up chasing our tails as the patient steadily drowns. This cannot be an appropriate treatment modality.

      If the problem is peripheral vascular collapse due to severe vasodilation, then it makes sense to vasoconstrict the patient early. This does not simply mean using a vasopressor however, as the majority of patients with septic shock have significantly impaired myocardial contractility. If we simply increase afterload using a vasopressor, then the left ventricle will dilate and fail. We need to ensure that the ventricle has sufficient inotropy to cope with the increased afterload that it will be subjected to before we vasoconstrict the arterial tree. That means we need to know the inotropy state of the heart. Unfortunately, ejection fraction is dangerously misleading in this situation, frequently appearing to be very high (85% is quite common) due to the very low afterload. This is often interpreted as indicating good cardiac function when in fact it simply means that the heart has nothing to stop it ejecting! Treatment based on ejection fraction in sepsis is profoundly dangerous.

      • Grant Price says

        Hit em hard and fast and then wait is a good lesson. Without doubt too much fluid ends up being given to patients especially after the first 12 to 24 hours of resuscitation to treat hypotension or poorly specific or sensitive signs such as urine output. This is because proper assessment of fluid responsiveness takes more effort than just writing up another bag of fluid- shouldn’t be that way but it often is.
        I am going to take issue with the assertion above that treatment of impaired cardiac output in sepsis with synthetic inotropes is a useful strategy. This assertion is often based on the presumption that increasing cardiac output in septic shock is useful, by which I mean (and your patient hopes) that it will increase the chance that they will survive. We know that doing it late and to higher levels than normal (whatever normal is) kills you. There is a continually repeated erroneous fact that noradrenaline is simply a vasopressor- it is not- it would not make sense for it to be so from an evolutionary standpoint and in clinical practice if one chooses to measure cardiac output starting noradrenaline increases cardiac output- it is our own pre-emininent physiological inopressor. The only positive evidence in the literature of dobutamine being beneficial when used early is in Dr Rivers paper- which was single centre and used management strategies for sepsis management that had been discredited in the mid 1990s. There is also the inconvenient fact that in many studies of sepsis those patients with the highest cardiac outputs did worse. It may well be that the multifactorial cause of impaired myocardial function in sepsis is protective and that making the heart work harder is counterproductive. I don’t measure cardiac output on my ICU, except as a diagnostic tool and then doing a snapshot with bedside echo simply because people like to chase numbers and unfortunately the treatment of such “abnormality” with adding in dobutamine has not proven itself to be beneficial in terms of mortality in sepsis, but certainly has caused harm.

        No conflicts of interest. I am a Critical Care Consultant from Edinburgh, Scotland

        • says


          Will go over this just a bit here and then in more detail on the audio response.

          Not sure where all of the controversy comes from with dobutamine. At least in the states, it is dirt cheap and there actually is literature:
          [1] De Backer D, Creteur J, Dubois MJ, Sakr Y, Koch M, Verdant CVincent JL, The effects of dobutamine on microcirculatory alterations in patients with septic shock are independent of its systemic effects, Crit Care Med, 2006, 34(2):403-408.
          [2] Enrico C, Kanoore Edul VS, Vazquez AR, Pein MC, Perez de la Hoz RA, Ince CDubin A, Systemic and microcirculatory effects of dobutamine in patients with septic shock, J Crit Care, 2012, 27(6):630-638.
          [3] Gutierrez G, Clark C, Brown SD, Price K, Ortiz LNelson C, Effect of dobutamine on oxygen consumption and gastric mucosal pH in septic patients, Am J Respir Crit Care Med, 1994, 150(2):324-329.
          [4] Jellema WT, Groeneveld AB, Wesseling KH, Thijs LG, Westerhof Nvan Lieshout JJ, Heterogeneity and prediction of hemodynamic responses to dobutamine in patients with septic shock, Crit Care Med, 2006, 34(9):2392-2398.
          [5] Levy B, Bollaert PE, Lucchelli JP, Sadoune LO, Nace LLarcan A, Dobutamine improves the adequacy of gastric mucosal perfusion in epinephrine-treated septic shock, Crit Care Med, 1997, 25(10):1649-1654.
          [6] Martin C, Viviand X, Arnaud S, Vialet RRougnon T, Effects of norepinephrine plus dobutamine or norepinephrine alone on left ventricular performance of septic shock patients, Crit Care Med, 1999, 27(9):1708-1713.

          No patient-important outcome data (Just like all of the vasopressors including norepi)

  21. roslan says

    After reading and listening to all of you….i go back and read FEAST study mentioned by dr Marik and found out some interesting tidbit…when I first browse through it , I sort of dismiss it because it is a paediatric study in Africa…however its hard to ignore its result since methadologically its such a good study (randomised althoughbnot blinded, multicenter ,3000+ pt with multi country , exclude malnourished and gastroenteritis)

    Halfway through the study they found out that the mortality difference btw the group was minimal….so they amend the study protocol by increasing fluid bolus ( 20ml/kg to 40 ml/kg)….only then the mortality start to increase even more in bolus group….

    Then they analysed why pt died in bolus group by independent committee who is blinded to the groups….they found out bolus group has minimal pt who died with pulmonary edema contrary to the expectation…….

    This sort of show us too much fluid is no good and we are not yet able to understand fully why this group of pt died after bolus administration…..reperfusion injury after bolus have been postulated….

    And mortality in saline group vs album on group was similar just like on SAFE..


  22. says

    Amazing talk. Marik does a wonderful job of turning convention on its head. Is there enough here to change practice? I don’t think so.
    The fluid administration issues brought up by others is important. Patients who were sicker likely got more fluids in many non RDCT studies.
    My concern is using the FEAST trial to generalize to the rest of sepsis care. FEAST was in Africa and there were a lot of kids with malaria in the study. Not really applicable to my septic adult population. Also, kids are relatively invulnerable in comparison to adults (at least those that I see).
    From the standpoint of assessing volume tolerance or responsiveness, I don’t have NICOMs in my department but I’m glad to hear Marik plug passive leg raise since it’s free and easy.
    The concept here is really important. There should be a real RDCT going forward looking at the question.

    • roslan says

      Yes you are correct..in fact 57% of FEAST subject got malaria….so it still has about 1500 plus pt who is not malaria….and I think it still is the largest ever rct comparing bolus with no bolus….and talking about malaria sepsis ….there are a few small study showing increase mortality in bolus group in severe malaria before FEAST….it is postulated because malaria sepsis has different pathophysiology than your normal sepsis….they have a systemic microcirculation blockage causing MODS….Another case in point is dengue shock syndrome where you cannot give too much fluid because of predominant capillary leakage……i think it is enough to warn us not to generalise all sepsis are the same…..personalised approach is important so we need some help to achieve it, be it USCOM, NICOM or whatever……

      Another point to make since this study is done in Africa , the amount of bolus fluid is conservative compared to standard paed European and american guidelines…this is because there have minimal ventilator resource…..so the result is even more impressive…bear in mind they need to stop halfway because its too unethical to continue giving more bolus…and they are using heybittle peto statistics thatvis more robust and difficult to satisfy compared to other method…..and lastly I doubt we will ever see the adult trial in other country because of the ethical issue since fluid bolus have been adopted as standard of care….TQ

  23. rob phillips says

    Fluid is not as difficult as it sounds. Fluid responsiveness is an SV increment associated with fluid administration. Measure the SV with high fidelity beat to beat SV technology; if the SVI10-15%, FTc<350ms, 1<SMII10% give fluid. Monitor till the PLR or bolus dSV<5% then stop. This doesn't need an RCT it is basic physiology.
    Perhaps the solution is appointing "fluid/haemodynamic specialists" whose principle job is to identify fluid responsiveness and administer and monitor fluid using high fidelity SV measures. Fluid is the most common critical care intervention yet approximately 50% of all administered fluids are unnecessary or ineffective, and worse may be associated with iatrogenic failure, increased mortality, hospitalisation and costs. The cost-effectiveness of "fluid/haemodynamic specialists" would be proven quite rapidly I suspect and outcomes improved dramatically. Like much of the clinical care its not the science but the implementation which is most challenging.

  24. rob phillips says

    Fluid is not as difficult as it sounds. Fluid responsiveness is an SV increment associated with fluid administration. Measure the SV with high fidelity beat to beat SV technology; if the SVI10-15%, FTc<350ms, 1<SMII10%, give fluid. Monitor till the PLR or bolus dSV<5% then stop. This doesn't need an RCT it is basic physiology.
    Perhaps the solution is appointing "fluid/haemodynamic specialists" whose principle job is to identify fluid responsiveness and administer and monitor fluid using high fidelity SV measures. Fluid is the most common critical care intervention yet approximately 50% of all administered fluids are unnecessary or ineffective, and worse may be associated with iatrogenic failure, increased mortality, hospitalisation and costs. The cost-effectiveness of "fluid/haemodynamic specialists" would be proven quite rapidly I suspect and outcomes improved dramatically. Like much of the clinical care its not the science but the implementation which is most challenging.

  25. rob phillips says

    Fluid is not as difficult as it sounds. Fluid responsiveness is an SV increment associated with fluid administration. Measure the SV with high fidelity beat to beat SV technology; if
    1<SMII10%, give fluid.
    Then monitor till the PLR or bolus dSV<5% then stop. This doesn't need an RCT it is basic physiology.
    Perhaps the solution is appointing "fluid/haemodynamic specialists" whose principle job is to identify fluid responsiveness and administer and monitor fluid using high fidelity SV measures. Fluid is the most common critical care intervention yet approximately 50% of all administered fluids are unnecessary or ineffective, and worse may be associated with iatrogenic failure, increased mortality, hospitalisation and costs. The cost-effectiveness of "fluid/haemodynamic specialists" would be proven quite rapidly I suspect and outcomes improved dramatically. Like much of the clinical care its not the science but the implementation which is most challenging.
    (Sorry the values don't seem to be converting)

    • says

      Ok, if you wanted to give one article that discusses the parameters you mention above and their use for fluid assessment, which would it be?

      • Francesco Savelli says

        Sorry if I’m intruding into the discussion but as you may have guessed I’m very interested in this topic. I try to provide a bit of literature and explanations, then Rob could correct and be more exhaustive. I’ve found two studies in which the FTc (the time of ventricular ejection corrected for heart rate) has demonstrated to be a predictor of fluid responsiveness:
        1) Sturgess DJ, Pascoe RL, Scalia G, Venkatesh B. A comparison of transcutaneous Doppler corrected flow time, b-type natriuretic peptide and central venous pressure as predictors of fluid responsiveness in septic shock: a preliminary evaluation. Anaesth Intensive Care. 2010 Mar;38(2):336-41.
        2) Lee JH, Kim JT, Yoon SZ, Lim YJ, Jeon Y, Bahk JH, Kim CS. Evaluation of corrected flow time in oesophageal Doppler as a predictor of fluid responsiveness. Br J Anaesth. 2007 Sep;99(3):343-8. Epub 2007 Jul 9.
        The first study has been made with the USCOM and the second with the oesophageal doppler, the results are similar with a cut off value of 350-357 (If the FTc is lower high probability of fluid responsiveness). But there are also studies in which it failed to reach a significant difference. For example, in the study of Thiel about passive leg raising, performed with the USCOM, the difference between responders and non responders was not significant, though very close to it.
        Steven W Thiel, Marin H Kollef and Warren Isakow. Non-invasive stroke volume measurement and passive leg raising predict volume responsiveness in medical ICU patients: an observational cohort study. Critical Care 2009, 13:R111

        FTc is considered to be a preload index, obviously the greater the left ventricular preload, the greater will be the time required for blood ejection. The problem is that FTc is also influenced by inotropy and afterload (if systemic resistances are very low cardiac empting will be quicker). It’s clear that the evaluation of the FTc alone could be measleading and obviously it’s not reliable if atrial fibrillation is present. And here, in my opinion, it comes the genius of Prof. Brendan Smith who has created and validated a formula for the determination of inotropy (which includes also the flow time).

        Brendan E. Smith,Veronica M. Madigan. Non-invasive method for rapid bedside estimation of inotropy: theory and preliminary clinical validation. Br J Anaesth. 2013 Oct;111(4):580-8.

        He has studied patients admitted to the emergency departement for non cardiogenic shock. First he has found that many patients in the sepsis group had cardiac disfunction (assessed by a low inotropy index), and he has also clearly found that a value of inotropy index lower than 1,2 predicts a negative response to fluid loading.
        Smith BE, Madigan VM. Inotropy Index Accurately Predicts Fluid Responsiveness In Volume Resuscitation. Emergency Medicine Australasia (2009) 21 (Suppl. 1)

        The problem is how to decide to stop fluid loading in patients with good inotropy: if they have a low stroke volume index (< 35 ml/m2) it seems logical to go on until the SVI reaches the normal range, if SVI is already higher I think we have to go on with fluid challenges or passive leg raising test, always evaluating the response of cardiac output. But probably if the inotropy is good, SVI is in the normal range, but the patient is still hypotensive and hypoperfused it's time to start a vasopressor.

  26. says

    Wow, All I can say is Wow! Never thought this would blow up like this (yes, I did). My response to all of this can no longer be a wee. You all will have to wait one more week for my response, which will be the next podcast.


    • Nikolay Yusupov says

      I hope in the next weeks podcasts you will also touch upon prehospital sepsis management and fluid administration. Given the NYC Remac proposed this Sepsis protocol, do not get surprised when SNF patients with ams start showing up in your ED with 2 large bore IV’s and 2 Liters of fluid on board when they hit your ED.



      1. Begin Basic Life Support Shock Measures.
      2. If the patient is demonstrating signs of inadequate ventilation, perform Advanced Airway Management*.
      3. Begin rapid IV/Saline Lock infusion of Normal Saline (0.9% NS) or Ringers’ Lactate (RL) via one to two large bore
      (14-18) gauge catheters, up to 2 liters, via a macro-drip. Attempt IV access no more than twice. Consider using
      the intraosseus route if peripheral attempts have failed.
      a. Accurate documentation of pre-arrival fluid administration is required.
      4. Begin Cardiac Monitoring, record and evaluate EKG rhythm.
      5. Measure and record lactate level (if available).
      6. Measure and record oral temperature (if available), also consider using last temperature obtained at patient’s
      facility (if available).
      7. Transport decision.

  27. Graeme pickford says

    Back when I was young, we used to try and drown our trauma patients in normal saline too. I recall going to a Ken Mattox lecture during the time when all the ATLS instructors and luminaries were still putting up some heated resistance; and he was talking about the third of hypotensive trauma patients who got better whatever you did, including a saltwater drowning– “if you put a towel clip on their sternum, they’d still got better….this doesn’t necessarily mean that it was good for them!”
    I propose that the concept of ” fluid tolerance” is the towelclip of sepsis.
    Gentlemen (and ladies) we have the technology…….. (And the science).

  28. Lorne says

    I’m intrigued by that statement Dr. Marik makes by saying that dexmedotomidine potentiates norepinephrine’s pressor effect. Anyone aware of data to support this? I can’t find much.

    • Paul Marik says

      In an experimental sepsis model dexmedetomidine (an ?-2 agonist) was demonstrated to increase the pressor responsive or norepinephrine.[1] In the subgroup of patients with sepsis in the “Maximizing Efficacy of Targeted Sedation and Reducing Neurological Dysfunction trial”(MENDS), patients who received dexmedetomidine had a lower pressor requirements, more delirium- and coma free days and a better survival than patients who received lorazepam.[2]


      1. Geloen A, Chapelier K, Cividjian A et al. Clonidine and dexmedetomidine increase the pressor response to norepinephrine in experimental sepsis: A pilot study. Crit Care Med 2013; 41.
      2. Pandharipande PP, Sanders RD, Girard TD et al. Effect of dexmedetomidine versus lorazepam on outcome in patients with sepsis: an a priori-designed analysis of the MENDS randomized controlled trial. Crit Care 2010; 14:R38.

  29. Erik Kistler says

    Thanks Scott, for the talk by Paul Marik. Definitely interesting. My take is that Dr. Marik is a bit of a conspiracy theorist, with just enough veracity to keep you listening. My head grew numb at the number of inconsistencies, inaccuracies, and just plain unsubstantiated opinion that was said. However, he does have several points which I think merit consideration: the first is that sepsis happens at the microcirculation, and it is there that we should be focusing. That we are currently unable to effectively intervene here except in a very gross manner by manipulating macro values of flow, fluid volume, etc., should not take away from the fact that microcirculatory perfusion is really what’s important in sepsis, and Marik should be commended for pointing this out.
    The other point that he got correct was that sepsis is not, strictly speaking, a disease of inadequate oxygenation to the microcirculation, but perhaps of utilization instead. I’m not entirely clear that it has anything to do with oxygenation at all, as it appears to me that improved outcomes with oxygenation have more to do with improved flow than anything else.
    Finally, I’m not sure I buy the diatribe against GDT, but Marik points out, obliquely, that there is a real lack of understanding on a fundamental level about the “goals” of GDT and how to best achieve them. A lack of real science is now being obscured by the marketing of devices that are being used to “optimize” (usually “maximize”) various macro-hemodynamics without a clear understanding of the physiology involved. Although it’s gratifying to see that each new device seems to improve outcomes, I’m not sure this approach will lead to optimization of care without a more complete understanding of the pathology.

  30. David J. Weber, Esq., FP-C says

    Great! Now I have to re-read all of the cited material along with the numerous newly revealed (to me) works to assure a philosophically un-biased pathophysiologically current lesson.

    My only conflict is that I follow orders, which is more personal than professional.

    Thanks Doc’s.

  31. Craig Rosebrock says

    Love the podcast. It really is refreshing that folks challenge the dogma. When I was in medical school I used to spend hours in a small campus library hidden under a set of trees …. Missed by most of my collegues because it was so small. It has books and assorted things of medical history and most were well over 200 years old. I recall one day reading a book on respiratory disorders from a series of lectures at a Berlin conference. I recall it being written in the late 1800’s. The treatment for asthma was ether or chlorophorme. Also what was interesting was the five page long physical exams and the fascinating explanation of pathophysiology. These scientist physicians knew what was wrong on an outstanding level of detail. What struck me is that each patient was a challenge and grouping folks into a syndrome or disease was hard because the literature had yet to formally placed nomenclature on disorders. Also how many of us did a really experiment in medical school?

    My point is that we all too often clump folks into sepsis, heart failure, liver failure, this or that and expect to treat them the same based on evidence for their syndrome or diagnosis. I am a locums doc and travel over seven states practicing critical care and pulmonary medicine and see a lot of protocols for this and that. The sepsis protocols are all a bit different but roughly follow Rivers’ approach. I have yet to see more sophisticated gadgets such as socom. I have learned as most of us here that each patient is unique. No septic patient is the same just as no asthmatic is the same. In today’s rat race of medicine we have lost some of our art. Some busy docs I know would just assume spend 2 minutes to fill out preprinted order sets and go to the next person. Our lives are dictated by rvus and administrative time constraints. Seeing the patient at the bedside and methodically working thru their issues is a luxury now a days. If not careful we end up only giving real time to those that vere off the expected path we have in our minds eye. My patient is not getting better with the usual sepsis cocktail…yikes…I better go see them again.

    I think the podcast is important for a number of reasons. Dr Marik show courage to buck the proven dogma that is only been time tested for a little over a decade. I think critically ill patients need very keen eyes and thought every step of the way. We can’t plug these folks into a protocol and hope for the best. I think protocols are great when you use them as a guide but are brave enough to do what most of the docs commenting here do..which is THINK about why we do what we are doing.

    Regarding data and gadgets, they are all useful in my opinion. However my understanding of fluid dynamics is that the calculations and theory is difficult to translate into humans. Fluid mechanical calculations are usually for fluids in fixed resistance and compliance vessels. The circulatory system In humans seems a bit complex to directly apply mechanics for non natural structures.

  32. says

    I have followed the recent discussion of resuscitation in sepsis with extreme interest – mortality in sepsis remains high despite significant progress by leading intensive care physicians from the Rivers study through the current Surviving Sepsis Campaign. I applaud these efforts which have already reduced mortality in sepsis, but we should not be satisfied with current outcomes for septic patients. Dr. Paul Marik (Chest, 2014, to appear) cites the key goal of resuscitation as adequate cardiac output (CI > 2.5) at adequate pressure for perfusion (MAP > 65 mmHg – recall that a blood pressure of 110/70 – regarded by many as ideal in a normal adult, corresponds to MAP = 83 mmHg!) – I strongly agree – and argues for a combination of fluids and pressors. Let me add that I have enjoyed many interesting, insightful and informative conversations with Dr. Marik, and I have learned much about the importance of understanding and applying the physiology of hemodynamics and perfusion.

    So, how do we get there – adequate flow at adequate pressure? Monitoring flow and pressure is not enough, there is more to the dynamical state of the cardio-vascular system. For example, both RV and LV dysfunction are common in sepsis, c.f. a recent review by Dr. Antoine Vieillard-Baron (Septic cardiomyopathy. Ann Intensive Care 2011;1:6). Here is where point of care echocardiography comes into play (c.f. Vieillard-Baron, A., et al. Bedside echocardiographic evaluation of hemodynamics in sepsis: is a qualitative evaluation sufficient?). Intensive care medicine, 2006;32: 1547-1552). Point of care echo can not only guide resuscitation; it can identify fundamental problems such as RV dysfunction. For this reason, Greenhalgh and Patrick called for the development of an indwelling TEE probe for managing hemodynamic instability (Perioperative transoesophageal echocardiography: past, present & future. Anaesthesia 2012;67:343-6).

    We at ImaCor agree – and for this reason we developed a miniaturized TEE probe, FDA-approved and CE Marked to remain indwelling, for 72 hours (hTEE, ImaCor Inc., Garden City, NY, USA) – with four studies in high impact journals this year. Although an appropriate study is needed to assess its utility in managing septic patients, early studies in other areas have proved promising (c.f. Vieillard-Baron A, et al. A pilot study on safety and clinical utility of a single-use 72-hour indwelling transesophageal echocardiography probe. Intensive Care Med. 2013;39:629-35.)

    Let’s regard important progress to date in resuscitating sepsis patients as an excellent first step and let’s continue to look for improvements in management which lead to improvements in outcome.

    Harold M Hastings
    Co-Founder and CTO, ImaCor Inc.
    December 13, 2013

  33. Gavin denton CCP uk says

    I’m very interested in the utility of USCOM. However, so much of the data out there seems to have the companies sticky fingers on it. Thanks Scot for making the disclosure and conflict of interest clear from certain contributors in this discussion. Very interested to hear more about the BUSH protocol, but again there seems to be a tight association, with the company on this. Even if this is not the case, maybe USCOM should back off a little from commenting on the study and allow us to scrutinise the methodology and results when they are published. My opinion on the study will now already be guarded given the lack of up-front disclosure from uscom reps, even if in reality the study is independent of them.

  34. Dan Grove, pulmonary/CCM in Baltimore says

    Please ponder the following two logical deductions:

    1. High blood sugar is associated with bad outcomes in critical patients
    Therefore all our patients should have their blood sugars strictly controlled.

    2. Large volume fluid boluses are associated with bad outcomes in critical patients.
    Therefore we should have the volume given strictly controlled.

    It seems to me the two cases are analogous. Just as we should have been more cautious in applying the first without adequate rigorous investigation (i.e. more than one RCT) so too we should be cautious with the second. Mortality in sepsis is dropping. So whatever we’re doing is working. Let’s be careful.


  1. […] is hard to go past a Paul Marik talk on ‘Fluids in Sepsis’ as a candidate for ‘Ripper of the Week’. We all owe a debt to Scott Weingart for sharing this […]

Speak Your Mind (Along with your name, job, and affiliation)