ProMISe Trial for Sepsis: Usual Care 3, Goal-Directed Therapy 0*
* (since 2014)
The ARISE (Australasia) and ProCESS (U.S.) trials, published in 2014, each demonstrated no advantage of protocolized care for sepsis over conscientious usual care. For those remaining unconvinced, the U.K.-based ProMISe trial is available in the New England Journal of Medicine.
ProMISe extends the growing global footprint of what some will call the refutation of sepsis protocols based on the original Rivers trial, and others will consider proof of their efficacy by their absorption into standard care.
ProMISe enrolled 1,251 people with severe sepsis or septic shock at 56 centers in the United Kingdom. They were randomized to usual care (whatever the physician felt best) or algorithm-driven early goal-directed therapy which included continuous central venous oxygenation monitoring (using the original EGDT protocol from Rivers et al).
In the intervention group, 92% received a central line and arterial line as soon as possible, followed by 6 hours of continuous ScVO2 monitoring. In the usual care group, half of patients never received a central line, and central venous oxygenation was not measured in those who did.
As should surprise no one by now, there was no difference in 90-day mortality: 29% of patients died in both arms. The relative risk was 1.01 (CI 0.85-1.20). There were no meaningful differences in secondary outcomes. The numerical equipoise made the probability of more than a 15% relative risk reduction with protocolized goal-directed therapy very low.
A Brief History of Sepsis Trials
Like any historical text, the sepsis literature needs to be evaluated in context of the era in which it was created. When Rivers published EGDT in 2001, conditions were such that a lone single-center trial controlled by one primary investigator could change the standard of care everywhere. Critical care research and clinical research generally have evolved dramatically since then; so has our collective understanding of the fallibility of the endeavor, and thus (rightfully) our skepticism of the Rivers study.
Further, sepsis care itself has improved since 2001, probably due to earlier recognition and provision of fluid resuscitation, antibiotics, and vasopressors. Although the EGDT protocol wasn't necessary to achieve this, the Rivers trial was a crucial catalyst (among others) for the social, pedagogical, and organizational sea changes required for sepsis surveillance and care to receive the appropriate priority and resources at medical centers around the world.
Are Sepsis Protocols Dead? What's the Difference?
Proponents of classic sepsis protocols surely believe the entire protocolized bundle is unnecessary most, if not all the time. But they rightly worry about the weakest links in the health care chain: lower-performing physicians, midlevels and nurses who (it is feared) are looking for any excuse not to place a central line or vigilantly follow septic patients' clinical status.
With a wink and a nod, I imagine subject-matter-expert-thought-leaders say to each other, we realize treating "goals" like CVP, ScVO2, lactate, and hemoglobin have no effect on a septic patient's outcome -- but pressuring community hospital-dwellers to repeatedly check them all will at least get them to the patient's bedside, make them use their brains, and force communication among team members.
Unfortunately, it's a compelling argument. Sepsis protocols probably have improved care overall in just such a way, and this is why proponents get upset when the idea of protocols themselves is challenged.
But in ARISE, ProMISe and ProCESS, the intervention-arm patients received more of all the "goal"-directed interventions (central lines, dobutamine, vasopressors, red blood cells, etc): it just didn't make any difference in their outcomes. All three trials' usual-care arms either forbade the use of central venous oxygenation monitoring, or physicians chose to use it very sparingly. This central variable of the EGDT protocol is apparently not needed at all -- which means dobutamine and red blood cells should not be protocolized.
What all the patients in all the trials had in common was early receipt of antibiotics, fluids and vasopressors, and a high survival rate from severe sepsis and septic shock.
Although well-intentioned, continuing to advise wasteful busywork simply to promote general conscientious care (after the crucial therapies have been provided) now seems a bit outdated, if not cynical. The consistent signal is, we must identify sepsis early and get antibiotics, fluids, and vasopressors started a.s.a.p. An algorithm for the next 6 hours is simply not needed to achieve that.
The Post-EGDT-Protocol Era Begins
It's time for higher-performing centers to lead the new era of post-EGDT protocol sepsis care. Early goal-directed therapy using the classic protocol wastes resources, the most precious of which is attention. Chasing labs like ScVO2 doesn't help patients, but does consume cognitive energy and time that should be spent elsewhere to greater effect in the ICU.
This part won't be hard: I suspect that some major academic centers are already shredding their reams of printed early-goal-directed therapy protocols, or informally considering most of the algorithm optional.
The bigger challenge will be for professional societies to issue new or amended guidelines responding to the game-changing evidence of 2014-2015 in a rational and responsible way. These new guidelines should eliminate the debunked portions of the EGDT algorithm and "goals" in favor of a simplified sepsis treatment pathway:
Identify patients with sepsis using accepted screening tools.
For patients with sepsis, collect cultures and give antibiotics, fast.
For severe sepsis, infuse ~2 liters crystalloid over a short period of time.
If hypotension persists, start vasopressors (through a central venous catheter and guided by an arterial catheter as soon as feasible).
Titrate vasopressors to achieve a mean arterial pressure ≥ 65 mm Hg (with a higher or lower target for individual patients).
That's it. Of course physicians can continue to do whatever else they want (checking CVP, ScVO2, lactate, etc., and responding however), especially in more severely ill patients with sepsis, who were a minority of patients in these trials. But no such maneuvers should be advised as standard practice, as there is no detectable effect on outcomes, in the aggregate.
The fears of a sepsis-protocol-free planet are ironic, because simplifying the treatment of severe sepsis should improve outcomes by making it easier to diagnose and treat sepsis, which is the whole point. First on the agenda: eliminating the advisement for immediate placement of a central venous catheter. My anecdote-based suspicion is that under the current regime, all-too-human emergency medicine physicians sometimes unconsciously hesitate to make the diagnosis of sepsis, knowing it means they will be sidelined in a busy E.D. to place a central line (which, their clinical experience rightly tells them, is often not needed). Simplifying the sepsis pathway would change this psychology, allowing faster diagnosis and treatment with antibiotics.
One challenge will be reaching consensus on how long to safely observe a hypotensive patient receiving crystalloid before a central line is considered necessary, and whether starting vasopressors through a peripheral IV briefly during that time is safe or advisable. That said, the physicians in ARISE, ProCESS and ProMISe seemed to navigate these waters without harm.
Call the new care standard a sepsis pathway, call it a simplified sepsis protocol, call it something else, but let's at least call it evidence-based. The current sepsis treatment guidelines can no longer be called that.