Early goal directed therapy does not improve outcomes in septic shock (ProCESS) - PulmCCM
Mar 212014

Can we finally “Just Say No” to the mandatory use of central venous catheters and central venous saturation in severe sepsis and septic shock?

by Muhammad Adrish, MD

In a single center study published in 2001, Rivers et al reported that patients with severe sepsis and septic shock had significantly lower mortality (30.5% vs 46.5%) when a 6-hour protocol of early goal-directed therapy (EGDT) was instituted. The protocol for EGDT called for placement of a central venous catheter (CVC) and monitoring of central venous saturation (ScVO2) and central venous pressure (CVP) to guide use of intravenous fluids (IVF), vasopressors, inotropes and packed red-cell (PRBC) transfusion. Trigger points were recommended for these interventions: crystalloids or colloids were recommended if CVP <8; vasoactive agents when CVP goals were met but mean arterial pressure (MAP) remained <65 mm Hg; and ScVO2 of < 70% triggered transfusion of packed red blood cells to target hematocrit of 30% (or the use of inotropic agents like dobutamine, if hematocrit was already >30%).

Early goal directed therapy rapidly became standard care for patients with severe sepsis; protocols were implemented at hospitals around the world incorporating all elements of the care bundle, but it was never clear which interventions were responsible for the observed benefits.

Since then, various investigators have asked whether all elements of the standard EGDT protocol are necessary or helpful, in light of evidence that dobutamine does not improve microvascular perfusion in septic shock, transfusion to hematocrit of 30% appears harmful in critically ill patients, lactate measurements may be an adequate substitute for the invasively obtained ScVO2, and single center studies often show exaggerated or erroneous results.

In other words, a multicenter randomized trial to assess the efficacy of all elements of protocol-based care suggested by EGDT was long overdue. Big news: the landmark results of the ProCESS trial were reported in the March 18, 2014 New England Journal of Medicine.

What They Did

Investigators randomized 1341 patients with severe sepsis or septic shock (hypotension + SIRS criteria) at emergency departments in 31 U.S. hospitals to either:

  • Protocol based early goal-directed therapy emulating the Rivers protocol, with mandatory placement of a central line to continuously monitor ScVO2 and CVP, administration of IVF, vasopressors, dobutamine and PRBC (n=439),
  • Protocol-based standard therapy, defined as 6-hour protocol prompted resuscitation with administration of IVF till “clinical” euvolemia and PRBC transfusion to goal hemoglobin of 7.5 g/dl or more; CVC placement and ScVO2 measurements were not mandatory (n=446);
  • Usual care: bedside provider directed all care without any prompted protocol (n=456).

During the study:

  • APACHE II illness severity scores were ~21 in all groups.
  • Central venous catheters were placed in 93.2% of patients in the EGDT group, 56.5% in the protocol-based standard therapy group and 57.9% in the usual care group. Similarly, ScVO2 was monitored in 93.6% patients in EGDT group, 4.0% patients in protocol-based standard therapy and in 3.6% patients in usual care group. Arterial lines for blood pressure monitoring were not required.
  • Patients in the usual care group received the least amount of IVF during the first 6 hours (2.3 L in usual care vs 2.8 L in EGDT and 3.3 L in the protocol-based standard therapy group).
  • Patients in the EGDT group received more dobutamine and PRBC during first 6 hours than protocol-based standard therapy and usual care. (dobutamine use, 8.0% vs. 1.1% and 0.9%, respectively; packed red-cell transfusions, 14.4% vs. 8.3% and 7.5%, respectively).
  • Patients in both protocol-based groups received more vasopressors although use of antibiotics, glucocorticoids and activated protein C was same. All these differences disappeared between 6-72 hours.

It's critical to recognize sepsis early, and this imperative was incorporated into trial design: early treatment with antimicrobials (76% received antibiotics in approximately 3 hours after arrival to ED; 97% received in 6 hours after randomization).

What They Found

There were no differences observed in 60-day mortality (19-21%), 90-day mortality or 1-year mortality between groups. Protocol adherence was good (89.1% adherence in EGDT group and 95.6% adherence in protocol based standard therapy group). No differences were observed in secondary endpoints including cardiovascular failure, respiratory failure, hospital length of stay or discharge disposition; incidence of acute renal failure was higher in protocol-based standard therapy (6% vs. 3% in the other groups).


What It Means

In this multicenter randomized trial, use of central hemodynamics and oxygen saturation monitoring did not result in better outcomes than usual conscientious care not incorporating these interventions. Similarly, the study found no significant advantage of protocol-based resuscitation with respect to morbidity or mortality when compared to the clinical judgment of bedside treating physician.

This study confirms the most important elements in management of sepsis: early recognition, early administration of antibiotics, early adequate volume resuscitation using clinical parameters and avoiding over transfusion. If these essential aspects of care are in place, protocolized measurements of central hemodynamics and oxygen saturation apparently do not improve patient outcomes measurably.

PulmCCM Editorial:

This study was 5 times the size of Rivers et al in NEJM 2001, a single center trial inherently more prone to error and bias, and challenged by some as methodologically unsound. "Usual care" has progressed significantly since 2001 -- in part due to increased recognition brought by the Rivers trial. Although that trial created what now looks like a lot of unnecessary extra testing and invasive catheters (not to mention all the reams of paper to print those protocols), Rivers et al did prompt everyone to "think sepsis" and give antibiotics and fluids earlier, undoubtedly saving many lives.

The good news for practicing physicians is that we fought the EGDT protocol to a tie. It's not needed in most patients with severe sepsis, as long as we do our jobs well: acting early, then closely monitoring septic patients for perfusion and response to therapies. Eliminating the indication for dobutamine (likely useless) and transfusion triggers of Hct < 30% (possibly harmful) are surely steps out of the darkness as well.

More than half of the patients in the non-EGDT arms received central venous catheters, but since so few of them (4%) were used to monitor ScvO2, the ProCESS trial essentially proves that -- given good care otherwise -- serially measuring ScvO2 is not helpful in treating sepsis.

Some will point out that the patients in the original EGDT trial were somewhat sicker (true), but this is not a valid argument given that APACHE scores were roughly similar and there would have to be gross dissimilarities present to overcome the huge power advantages of ProCESS over Rivers et al. (When ProCESS authors restricted their analysis to the sickest patients, protocols still showed no benefit.) Besides, the question is not "whose trial represents truth?" -- it is "whose trial represents truth today?" -- and that answer seems manifestly clear.

All that said, protocols built a culture of vigilance for sepsis care: urgent action, frequent bedside examinations and therapy tweaks -- which probably improved usual care enough to produce the negative finding here. (The same phenomenon has possibly occurred with sedation interruptions.) It would be a tragedy in slow-motion if abandoning sepsis protocols led to a deterioration of usual care into apathetic care, losing the hard-won improvements in survival from severe sepsis and septic shock.

takeawayClinical Takeaway: Protocols don't improve survival in severe sepsis and septic shock, but especially in the golden early hours, they might still have value as a handy checklist to keep everyone on top of their game. Early antibiotics, adequate fluid resuscitation and vasopressor support are the essential components of care for severe sepsis and septic shock.

The ProCESS Investigators. A Randomized Trial of Protocol-Based Care for Early Septic Shock. N Engl J Med: DOI: 10.1056/NEJMoa1401602

Craig Lilly. The ProCESS Trial — A New Era of Sepsis Management: NEJM editorial.

Should lactate clearance replace SvO2 in sepsis protocols? (Pro/Con, CHEST)

Surviving Sepsis Guidelines: PulmCCM Review & Update

March 28, 2014: The Surviving Sepsis committee issued a statement responding to the results of the ProCESS trial, in which they “continue to recommend all elements of the current bundles”:


Liked this post? Get a weekly email update, and explore our library of clinical guidelines, practice updatesreview articles. and board review questions.

PulmCCM is an independent publication not affiliated with or endorsed by any other organization, society or journal referenced on the website. (Terms of Use)

Authors: contribute your work in a guest post.

  25 Responses to “Early goal directed therapy does not improve outcomes in septic shock (ProCESS)”

  1. The other danger highlighted by this trial is the premature institution of “Core Measures” or “Quality Measures” by CMS or payers that box everybody in and waste a lot of time, effort, and resources if later their underpinnings are refuted. We should be circumspect about forcing practitioners to start doing things based on the “latest evidence” given the track record of many of these things. Evidence is fluid and dynamic, not static.

    • Dr Lilly the NEJM editorialist made a related point: “State legislation and clinical guidelines, including those endorsed by the National Quality Forum, should be updated to remove the requirement for central hemodynamic monitoring and to focus on less costly, lower-risk, and equally effective alternatives.”

      And won’t the Surviving Sepsis Guidelines need revision as well?

  2. Like any other study, the results need to be interpreted with caution. This study never said that doing nothing is as good as EGDT. It just highlighted which elements of goal directed therapy will not improve outcome. For example early identification sepsis and prompt antibiotic administration is still paramount. Fluids; even in “usual care group” patient received over 2 litres IVF in the early period following identification which means many of these patients might still benefit from them. Routine CVP/SvO2 use has been abandoned by many intensivists already. At our institution we use it may be once or twice a month to decide the treatment especially when its more of a mixed shock picture. We still use it quite often in post cardiac surgery patients. Our intensivists have gotten so use to of doing bedside echocardiography (and it has been shown in many studies to be quite useful in management of any patient presenting with shock) that we use it way more often than CVP to decide our management.

    • “This study never said that doing nothing is as good as EGDT.”

      Agreed. So why title this piece “Early goal directed therapy does not improve outcomes in septic shock?” This title in pulmccm.org as well as the conclusion from the NEJM article stating that “protocol-based resuscitation of patients in whom septic shock was diagnosed in the emergency department did not improve outcomes” are both extremely misleading. What underpins these poorly-worded proclamations is that the Rivers trial was wrong; how inaccurate is that!

      A more accurate conclusion is that the type of protocol-based care does not change outcomes as long as patients have first received standard care in the first couple of hours (abx + IVF, as all pts in PRoCESS rec’d prior to randomization), something that Rivers patients (and for that matter, most septic patients outside of trials) were not routinely receiving in 2001.

  3. Great review – the only downside to this study is the risk that those who were reluctantly semi-following EGDT will now go back to their “usual care” which isn’t the usual care in this study… Just like some feel that 36 degrees is “normal” and that cooling is no longer required…


  4. Excellent review! It goes to show again the importance of a diligent and astute clinician. It goes to show again that sticking to the basics may provide better care and interventions need to be accessed for benefits. Doing less is doing more.

  5. Excellent analysis! This article again shows the importance of true medical practice and clinical research.

  6. Excellent review..couple of observations to add to this discussion.
    The usual-arm received lower amount of fluids on an average than the protocol based arms, yet had less incidence of AKI. They had a higher pH and and a lower chloride level at six hours. Gives us something else to think about- less chloride containing fluids might have translated into lower incidence of AKI in this study.

  7. Yes you are right Abhishek. For chloride containing fluids, evidence is there.

  8. Great PCCM editorial and takeaway statement.

  9. Dr Aberegg’s point cannot be overemphasized, given our current plunge into government controlled medical care.

  10. this study did not reach statistical significance when comparing mortality in the 3 groups separately, they did after they combined the 2 protocoled groups vs standard care group, which is not clearly defined, so this really does not conclude that central lines/ SVO2 are not helpful.
    also there is a higher percentage of lactate>5.3 in EGDT group “38 vs 26%”
    also the EGDT should not have included the blood transfusion part since more recent evidence points otherwise
    I take those results with suspicon I don’t think this is enough evidence to abandon the EGDT yet.

  11. Serum lactate levels were identical in all 3 groups. (Table 1. baseline characteristics). Blood transfusion to a goal Hb of 7.5 was used in “protocol based standard therapy group” (something that reflects current ICU practise) but was not targetted to Hct of 30 as suggested by EGDT but it still did not make any difference when compared to usual care group. And finally there were no differences observed in 60-day mortality, 90-day mortality or even 1-year mortality between all 3 groups.

    • 1-read the appendix lactate > 5.3 in EGDT group was 38% in the 2 other groups it was 26%.
      2- over transfusion in EGDT may have confounded the benefit of the rest of the EGDT protocol, since the new evidence supports lower HB that should have been use in both protocol groups.
      3- the mortality was not different but the P value was 0.8 , the P value was only significant when the 2 protocol groups were combined together.

  12. I have trouble understanding why neither the paper nor the various reviews on the topic do not elaborate on the renal failure requiring renal replacement therapy. This is a devastating complication for a patient and the incidence was double in the ProCESS arm. The ProCESS arm’s management was in fact, inferior at best.

    • I thought it most likely due to chance since it was not duplicated in the “usual care” non-protocolized arm (only 2.8%) and there was not a clear physiologic reason for why it occurred, in the reported data. For example, the usual care group received the least fluids in the first 6 hours (but had the lowest incidence of renal failure).

      • I would tend to disagree on this. The P-value was 0.04, which means that if was due to chance alone, this chance is 4%–which is very very low. Also, if we repeat the same trial 100 times in a theoretical realm, 96 times would have shown the same thing, which is renal failure in the ProCESS group. For me the usual care group is equivalent to physician judgement + the surviving sepsis guidelines. Not everything that has a physiological reason works (for example this trial showed that ScVO2 monitoring failed) and vice versa. I personally hypothesize that delaying a central venous access for treatment (not monitoring) is harmful.

        • But … people got central catheters at equal rates in the protocol (57%) and usual care (58%) arms. Your point about p values is statistically true, but if we compare enough secondary endpoints (20 on average) we will eventually get a spurious p value of 0.05 for one of them. The alternative is there was something else unmeasured here causing the renal failure, because (to me) the published results don’t provide a clue.

  13. I wonder if there is any way to know what usual care amounted to. I suspect that it amounted to very similar care. I suspect that many practitioners giving usual care made very frequent assessments in the first 6 and 12 hours. They likely used other physiologic assessments that have been validated like lactate clearance or dynamic ultrasound to guide fluids. Though they gave less fluid in six hours they still gave over two liters in the first six hours. This information does not seem to be in the appendeces but it is relevant because many institutions esp without closed units likely do not have the same ‘usual care’ as the ‘usual care’ provided. The outcome is as it was because in the 13 years since Rivers ‘usual care’ in many icus has supplanted it. I think in other icus this is not the case as per the intermountain sepsis trial.

  14. The Surviving Sepsis committee issued a statement responding to the results of the ProCESS trial, in which they “continue to recommend all elements of the current bundles”:


    • I think that is the wisest, since it “forces” those whose “standard care” would be sub-par to do at least their bundle. Those who have the know-how to do more now have the data to know that EGDT is not the be-all and end-all, as anyone can imagine. Is there anyone out there that thinks that 2014 medicine is the pinnacle, and that in 2050, we won’t have gotten a bit better? That simple principle mitigates against hanging on to current therapy for dear life. Gotta let go at least a little to evolve.


  15. Mohammad,

    Are you out of your freaking mind! I read the NEJM article and I laughed it off as a poorly constructed, conducted, envisioned, and executed study that never should have been published in the NEJM. I am embarrassed by my colleagues in critical care medicine who have not taken you to task for your outrageous conclusions. This study in no shape, way or form replicated the River’s study. IT ANSWERS nothing regarding Goal DIRECTED THERAPY which remains one the Cardinal Rules of Critical Care. It was not constructed to answer whether or not SVO2, early transfusion, early pressors or dobutamine improves outcome. The comparison group has no defined endpoints other than conventional care. What is that; some nebulous concept of resuscitation . Critical Care resuscitation has evolved greatly since the River’s article. CVP has all but been replaced by bedside intensivist echocardiography. No manuscript other than the River’s article methodologically tested early transfusion of blood products in critical illness. All other manuscripts have assessed transfusion in clinically stable patients whose Svo2 has normalized. All critical care physicians accept early antibiotics are key but only a component of the resuscitation bundle. Critical care bundles are for administrators, nurses, and house staff in training. They serve no other purpose than administrative.
    Essentially, the only thing that this study demonstrated was that observational outcomes across both experimental and control groups was no different. What does that mean? Absolutely nothing. This study does not advance the field critical care. It puts us back to the late 1990’s which I remember quite well. Any intensivist understands the power of the River’s paradigm irrespective of the methods you use to restore standard state conditions in critical illness.

  16. […] Early Goal Directed Therapy Dead? St Emlyn’s weighs in on the ProCESS Trial. Cliff Reid , PulmCCM and Academic Life in Emergency Medicine have also responded. [MG, […]