Sorry to bother you, but we have to talk about Rivers!


Early goal-directed therapy in the treatment of severe sepsis and septic shock, by Emanuel Rivers et al [1].


We all heard about it. We all read about it, hopefully even read it. We all have discussed it with friends over some beers. OK, maybe just me. The Rivers trial[1] was a landmark in critical care papers, and for better or worse, deeply changed how we managed sepsis in the beginning of 21st century. After that, a lot of papers were published studying how we can increase survival and quality of life of septic patients. In the last 2 years, the early goal directed therapy (EGDT) trilogy was published (ProMISe[2], ARISE[3] and PROCESS[4]), burying the way Rivers taught us on how to treat our patients. I’ll have to admit: the EGDT trilogy was the last nail in the Rivers’ coffin. But I have one big question: was the Rivers trial ever good enough to make all that fuss? How can a single center study was able to guide the way septic patients were treated during the last 15 years? Today, we’re gonna talk about what is between the lines, the pitfalls, and the reasons why we should not blindly believe in Rivers’ fantastic results.

The basics

So, it’s late 90s, and the concept we had about shock was: oxygen delivery/consumption imbalance, global tissue hypoxia, aerobic metabolism is shifted to anaerobic, and lactate production is increased. Did I really said this? Oh, fuck! Moving on. Physiologic endpoints, like vital signs, urinary output or CVP are not really good in detecting abnormalities in oxygen metabolism. It was thought that the assessment of central venous oxygen saturation (ScvO2), a surrogate for tissue oxygen extraction, and its normalization (if we can call anything normal in human physiology), could be a marker of better oxygenation and resolution of shock. Oh boy, if at least a expensive, fancy catheter designed to continuously monitor the ScvO2 existed. It would be perfect. Wait! There is one! And guess who invented it?

Inspired by the work of Shoemaker, who showed us that increasing delivery of oxygen in high risk surgical patients in the perioperative setting reduced mortality, Rivers created a similar protocol to treat the early stage of septic shock. It was a single center study, non blinded, at Henry Ford Hospital, Detroit. Patients were enrolled from 97 to 2000. Inclusion criteria were: adult patients, presence of 2 of 4 SIRS criteria AND systolic pressure ≤ 90mmHg after fluid resuscitation (20 to 30ml/Kg of crystalloid bolus) or lactate levels ≥ 4mmol/L. Exclusion criteria: pretty much everything else that was not sepsis. The study was conducted in a nine-bed unit in the emergency department by an emergency physician, two residents, and three nurses. Patients assigned to EGDT (intervention group) received a central venous catheter capable of continuous monitoring the ScvO2. Patients were treated according the EGDT protocol for at least six hours, and then, were transferred to the first available inpatient beds, where doctors were unaware of which group patients’ were enrolled.

The EGDT protocol: Fluid boluses were given to achieve a CVP of 8 to 12mmHg, after achieving this first step, if MAP < 65mmHg = vasopressor, if MAP >90mmHg = vasodilators. The target was a MAP between 65-90mmHg. After assured the MAP target, if ScvO2 < 70%, red cells transfusion until hematocrit ≥30%. If optimized, and ScvO2 still < 70%, dobutamine was infused (maximum dose of 20mcg/kg/min). It gets worse, if you couldn’t achieve those hemodynamic parameters, in order to decrease oxygen consumption, patients received mechanical ventilation, sedatives and paralytics. The control group: patients in the control group received standard care, which consisted in… Just a second, might be in the next page… Well, as you can see, nobody knows how resuscitation in control group was done.

Primary endpoint was in-hospital death. Secondary endpoints were resuscitation endpoints, organ-dysfunction scores, coagulation-related variables, administered treatments, and the consumption of health care resources. Assuming a rate of 10% refusal or exclusion, two-sided type 1 error of 5%, and an 80% power to detect a 15% difference in mortality, a sample size of 260 patients was needed. Kaplan-Meyer curve and risk ratio were used to describe the risk of death. A total of 263 patients were selected. Patients were around 65y/o, initial creatinine of 2.6mg/dL, lactate of 7mmol/L, they had a lot of comorbidities, and less than 10% were surgical patients. They were pretty sick, right? The APACHE II was around 21, the same as found in the PROCESS[4] trial (20.8).

More is…More???

In the first 6 hours, EGDT group received more fluids (4981 vs 3499ml), red cells (64% vs 18.5%) and dobutamine (13.7% vs 0.8%) than control group. Vasopressor use and mechanical ventilation were not different between groups in the early phase. Also, during the first 6 hours, EGDT group had higher MAP, ScvO2, and base deficit than control group. There were no changes in lactate levels. Interesting, isn’t it? Increasing the DO2 does not decrease lactate leves in the the early phase of septic shock!  Also in the secondary endpoints, EGDT group had less multi-organ dysfunction in the 7-72h period and less coagulopathy. The table below summarizes the mortality outcomes.

Amazing, right? For the primary outcome it gives us a NNT of 6.25! Couldn’t get better than this! But, there is always a but, to believe in those results, we also have to believe in the interventions made. In fact, it was not a single intervention, it was a bundle of care, with therapies guided based on hemodynamic goals, with the help of a new catheter waiting to be sold by Edward Lifesciences. It’s hard to know which specific intervention was responsible for the survival benefits, therefore, we’ll analyze them separately.

The use of CVP to guide fluid therapy: In 2008, Marik et al. published a meta-analysis[5] evaluating the role of CVP to guide fluid therapy. They found that CVP had an accuracy of 56% to predict fluid responsiveness. Coin flip, right? Now, we have lot of data associating positive fluid balance with increased mortality.

MAP >65mmHg: Physiologic studies showed that below this threshold there is a dramatic decrease in blood flow to the tissues. We don’t have any randomized controlled trial evaluating MAP targets below 65mmHg, and probably will never have. Here, 65mmHg was the MAP target for both groups( EGDT and control). But EGDT group had higher MAP than control during the study period.

Red blood cell transfusion: The TRISS trial [6], evaluated hemoglobin targets for septic shock patients, and concluded that there are no difference in mortality when comparing hemoglobin thresholds of 7g/dL vs 9g/dl.

Dobutamine: Most of ours septic patients already have a high flow state, with high cardiac output, especially in the early phase of septic shock. Hayes [7] (ICU Revisited link) showed an increase in mortality when you use dobutamine to achieve supranormal levels of cardiac output. So, that’s a NO!

Sedation and paralysis: To reduce oxygen consumption? Just don’t do it!

As you can see, today, we don’t have any evidence to support the use of those interventions alone, at least as a routine treatment of septic shock. I really don’t have a clue of what made this things work when used altogether. Probably the most heard critic of the paper is the high mortality rates in control group, which differs from other trials studying septic shock, even in the early 2000s. Maybe 46% mortality rate is too high for those living in developed countries, but, unfortunately, it’s a reality here in Brazil. The PROGRESS study[7], published in 2009, showed a 56% mortality rate in septic shock patients admitted to the ICU in Brazil.

The patent issue

With a quick search in, we found that the Presep’s patent, the catheter used in the study, was manufactured by Edward Lifesciences, and made by Dr Emanuel Rivers.

Rivers transferred the patent’s rights to the Henry Ford Hospital, before the study publication. It’s written in the article that Edward Lifesciences provided the catheters for the study. Now, that’s a huge conflict of interest, which was not published. It almost seems that they were trying to hide something.

Things get ugly

In the first paragraph of results section, the authors say that 288 patients were evaluated, and 8.4% (25 patients) were excluded or did not consent. Where were they? Were they in the EGDT group? Control group? Did they die? Why were they excluded? The article didn’t give any answer to these questions. But, we can find a clue in a Wall Street Journal issue (link), published in 2008, and we’ll quote: “Dr. Rivers, in an April interview describing how he proceeded, said “all 288 are randomized” — that is, placed into either the standard-therapy group or the other group — “and then some are excluded because they’re not considered severe septic shock” “…”Further muddying the waters, Dr. Rivers, despite having said all 288 patients were “randomized,” stated in a later email that “there were absolutely no patients pulled out before the analysis.” So, this is a serious charge. If you randomize, you analyze. Medical research for dummies, first lesson.

There is more: “Hospital statisticians who worked on it had concerns about the data and recalculated the results with all 288 patients, according to people familiar with the events. When they did, these people say, the death-rate gap wasn’t statistically significant.”…”Statisticians were especially concerned when they noticed that a relatively high proportion of the other 25 – those not included in the final analysis – were either conventional-therapy patients who survived or patients on aggressive therapy who died, say the people familiar with the events.”

Unfortunately, those “people familiar with the events” and “statisticians” were not named in the new, but still, it’s concerning. The guys who ran the study also had the catheter’s patent. Now you must be thinking: “Dear Lord, those Brazilians got really paranoid after all those political scandals in their country”. Yep, we did!

Final judgment

In the end, it a was single center study, non blinded, testing a bundle of care, which included a lot of thing, that when you isolate them, it seems most of them don’t work, and, in fact, some of them are harmful. Also, we have a non published disclosure, and methodological flaws that you don’t expect to find in a big journal like NEJM. Nor positive, neither negative. Let’s call it a draw!


1.     Rivers E, Nguyen B, Havstad S, et al. Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med. 2001;345(19):1368-1377.

2.     Mouncey PR, Osborn TM, Power GS, et al. Trial of early, goal-directed resuscitation for septic shock. N Engl J Med. 2015;372(14):1301-1311.

3.     Peake SL, Delaney A, Bailey M, et al. Goal-directed resuscitation for patients with early septic shock. N Engl J Med. 2014;371(16):1496-1506.

4.     Yealy DM, Kellum JA, Huang DT, et al. A randomized trial of protocol-based care for early septic shock. N Engl J Med. 2014;370(18):1683-1693.

5.     Marik PE, Baram M, Vahid B. Does central venous pressure predict fluid responsiveness? A systematic review of the literature and the tale of seven mares. Chest. 2008;134(1):172-178.

6.     Holst LB, Haase N, Wetterslev J, et al. Lower versus higher hemoglobin threshold for transfusion in septic shock. N Engl J Med. 2014;371(15):1381-1391.

7.     Beale R, Reinhart K, Brunkhorst FM, et al. Promoting Global Research Excellence in Severe Sepsis (PROGRESS): lessons from an international sepsis registry. Infection. 2009;37(3):222-232.ivers E, Nguyen B, Havstad S, et al. Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med. 2001;345(19):1368-1377.



Photo credit

J. Matt Buchanan


What's Your Reaction?

Confused Confused
Damn Damn
Geeky Geeky
Like Like
Scary Scary
Win Win
Angry Angry
Happy Happy

Sorry to bother you, but we have to talk about Rivers!

log in

Become a part of our community!

reset password

Back to
log in