Multicenter Trial Affirms Best Treatment for Severe Sepsis

Marrecca Fiore; Chad M. Cannon, MD; Emanuel P. Rivers, MD, MPH


October 16, 2012

Editor's Note:

In August, the Journal of Intensive Care Medicine published "The GENESIS Project (GENeralized Early Sepsis Intervention Strategies): A Multicenter Quality Improvement Collaborative.[1] The trial looked at whether the effectiveness of early goal-directed therapy (EGDT) or the 6-hour resuscitation bundle (RB) demonstrated in single-center trials could be re-created in a multicenter implementation involving both community and tertiary care hospitals.

The study used 2 strategies to assess treatment. The first included a before-and-after observational comparison of historical controls vs patients receiving the RB after implementation of early sepsis strategies in 4 community and 4 tertiary care hospitals. The second, conducted in 1 community and 2 tertiary care hospitals, examined sepsis patients not achieving all components of the RB vs patients achieving all components of the RB concurrently after GENESIS implementation.[1]

Among the findings of the study were that the RB, whether examined in a before-and-after or concurrent methodology, was associated with markedly lower mortality rates and a 5.1-day decrease in hospital length of stay.[1]

Medscape talked about the findings with the study's lead author, Chad Cannon, MD, and Emanuel Rivers, MD, who first championed early intervention strategies for sepsis back in 2001[2] and is considered the "father" of EGDT for sepsis.

Medscape: Dr. Cannon, as the lead author of this study, can you give me a little background on how the GENESIS Project came about?

Dr. Cannon: Sure. Back in 2001 when the original EGDT[2] results were published, the RB showed a tremendous reduction in sepsis mortality. However, there were a lot of hospitals that were kind of slow to respond to the literature. Gradually, the RB started gaining steam, especially after the first Surviving Sepsis Campaign released its recommendations in 2004. Most of the subsequent publications consisted of single-center before-and-after observations of benefit by applying the RB. What we wanted to do was a multicenter study looking at conditions before and after implementing these bundles and combine that with a look at the impact of full compliance with the RB.

Medscape: Dr. Rivers, do you see this trial as a continuation of your earlier work?

Dr. Rivers: Yes. The concept of the RB began years ago.[3] It was re-emphasized in the mid-1990s when we had a septic shock mortality of over 50% at [Henry Ford Hospital in Detroit, Michigan]. We decided to do a quality initiative, which ended up being a study of best practice. We created an algorithmic approach to managing sepsis in the earliest stages that, for the first time, included the emergency department (ED) because the waiting times were significantly long. We felt that sepsis management should mirror acute myocardial infarction, stroke, and trauma in order to realize the best outcomes. In 2001, we published in the New England Journal of Medicine showing that this approach to sepsis management results in a an absolute mortality reduction of 16%.[2] From that concept evolved, as [Dr. Cannon] said, many single-center studies, but we wanted to answer the question of whether the findings in this clinical trial in 2001 could be generalized throughout both large and small hospitals. So we did a multicenter collaborative study of 11 hospitals and pooled data. What we showed in this study[1] is a very similar mortality reduction to the original EGDT trial in 2001. But more importantly, it could be done in both community and tertiary care hospitals.

Medscape: Dr. Cannon, can you explain some of the findings of the GENESIS Project?

Dr. Cannon: The overall finding was that we saw a 14% absolute reduction in mortality associated with use of the RB. And, if you look at the 2 assessment strategies separately, this finding was consistent in the group after receiving the bundle vs the historical controls. When we looked at it from a compliance perspective in the concurrent groups, whether the RB was completely achieved vs not achieved, that was consistent as well, with a 15.3% reduction in mortality. This same mortality benefit was also seen in different hemodynamic categories, whether the patient was hypotensive, had a lactate level > 4 mmol/L, or a combination of the two. It didn't matter if these patients presented in the ED, on the hospital floor, or in the intensive care unit (ICU); we still saw a similar mortality benefit. As Dr. Rivers mentioned, we think it is unique that this study was done both in tertiary care/academic hospitals and community hospitals with similar mortality benefits at both types of facilities.

Medscape: Dr. Rivers, since your original study was published in the New England Journal of Medicine,[2] some researchers, many of whom were conducting single-center trials, have found it difficult to apply the early therapy and RB model in smaller hospital centers. Other researchers have found that not all of the RB elements may be equal. In other words, some elements of the RB are more important than others when it comes to reducing the risk for mortality from severe sepsis. Do you think this latest research will put to rest concerns about whether EGDT is appropriate for all hospital settings and re-emphasize the need to perform all elements of the RB?

Dr. Rivers: It is a good question. The medical community has to address the issue that these patients require a higher level of care than is currently given in order to improve outcomes. While this may be difficult for some, it is necessary for our patients. The RB is not only a scientific advancement but an organizational advancement; meaning that the ED, the hospital floors, and the ICUs were silos in regard to sepsis management. Prior to the EGDT study, there was no communication between specialties regarding the care of sepsis patients, and so here we have a scientific advancement in which we find that communication between specialties is key, and yet there is an implementation piece that was not present. Look at it in terms of other diseases: If you present with an acute myocardial infarction, a protocol is activated and coordinated definitive care is required in 90 minutes. When you look at any disease that has a high mortality rate, like strokes, heart attacks, and trauma, there is an organizational and communication plan in place. Sepsis has the highest hospital mortality rate of any in-hospital admission because you are 8 times more likely to die than with any other admission. It is the most expensive of all-cause hospitalization, costing $52 billion a year. Interestingly, for a disease as lethal and costly, you didn't have an organizational approach.

What we wanted to do is show, with the first study in 2001, that you could improve mortality if you treated these patients with the same early goal-directed approach that you used to treat other high-mortality diseases. The secondary issue comes in in terms of implementation. The hospitals were not geared to manage sepsis patients this way. If you look at the average [time to treatment] in this study, it was about 5.4-5.7 hours, which is consistent across the country and is actually a little more conservative, because with other institutions the average length of [time to treatment] can be 6-12 hours. We wanted to show that those hurdles to treatment could be overcome, and that was one of the primary goals of the study.

In addition, we wanted to show that a continuous quality improvement (CQI) process is important. When you look at those RB elements and you [doctors] say that you're doing them, but you don't actually do them, mortality can vary. And patient mortality improves by just accomplishing each bundle element. A CQI process provides feedback to the clinician to make sure all of those goals are being completed.

What we showed is that the time duration for each element of the bundle diminished significantly over the course of the study, which underlies the importance of what we are trying to accomplish. You can give antibiotics earlier. You can check blood cultures earlier. You can give fluids earlier. And you can accomplish these hemodynamic endpoints irrespective of what type of hospital center you are from.

Medscape: There has only been 1 FDA-approved drug therapy for sepsis, and that was taken off the market earlier this year. Why has it been so difficult to develop a drug therapy for treating this disease?

Dr. Rivers: I think the paradigm I spoke about earlier is amazingly similar. If a patient has a heart attack and sits in the ED for 6 hours, and then somebody calls a cardiologist and takes them to the catheterization lab, what do you think will happen to the patient? They either would die or the therapy wouldn't work. If you look at all the clinical trials [on sepsis drugs] in the past, they have given these drugs far too late in the natural course of the disease.[4] Patients presenting with sepsis frequently have symptoms for 24 hours before they come into the hospital.[5] What happens when you come to the hospital? Well, you are going to sit in somebody's ED for a few hours and then somebody will figure out that you are infected, and then you will end up -- we hope -- getting some standard care. A third of these patients or more are sent to the hospital floors to have a decompensation and then end up in an ICU later.

One of the beauties of this study is that we showed that you decrease the length of time or you decrease the amount of patients diagnosed on the floors. In this study we showed that if you implement a sepsis protocol, you actually diagnose your patients earlier, which results in a decrease of patients who end up on the floors because they are getting diagnosed in the ED and sent immediately to the ICU.

So, getting back to the original question, when you design a trial and you are giving a drug 18-24 hours after the onset of the disease, giving the disease a jumpstart, the drug can no longer pinpoint the disease the way it was meant to treat it.

When you look at drotrecogin alfa or recombinant protein C, the time to enrollment in the first clinical trial was over 17 hours after the onset of organ dysfunction and showed a mortality benefit.[6] In the second trial, while sepsis mortality has decreased 10%-12% over the last decade,[7] the time to drug administration was the same.[8] So, when you reintroduce the drug to test a patient population that's already seeing improved mortality rates, you are going to see a diminished drug benefit.

Thus, there is a graveyard for clinical trials of immunotherapy because of the delay in therapy and because of the failure to control for early diagnosis and management. Also if, in the future, time to enrollment in sepsis drug trials is decreased to 0-6 hours instead of 12-18 hours, I think you may find [a therapy that shows] some mortality benefit.

Medscape: How does a small community hospital go about bettering their organization so that they can start treating people earlier?

Dr. Rivers: There must be a standard operating procedure; you can't leave it to chance. Teach people how to detect infection early using blood cultures. Early administering of antibiotics [is necessary], and then when you get into the more advanced therapies such as maintaining central venous pressure and central venous oxygen saturation, you can actually coordinate with the intensivist. These community hospitals had community intensivists that would, basically, be called in when that patient came to the ED, and that ED many times would process the patient much quicker, get them to an ICU, and that intensivist would be called in earlier. Simply creating the standard operating procedure and allowing these patients to be rapidly advanced through the system alone has shown mortality reduction, and that can be accomplished in community centers.

These therapies are not beyond the capabilities of emergency physicians. It is just that some EDs may not be set up or geared to it, but there is nothing in this protocol that a board-certified emergency physician or intensivist should be unable to do.

Medscape: Dr. Cannon, what is the take-away from this study?

Dr. Cannon: I think it's to re-emphasize that these are not therapies that can only be done in a research setting or at a large academic hospital with the capabilities to implement them. Even smaller community hospitals can implement a standard operating procedure and follow that with a CQI effort to maintain compliance. It boils down to early detection of a high-risk illness and rapidly implementing a protocol that is going to decrease the time to checking a lactate, obtaining blood cultures, starting antibiotics, implementing a fluid challenge, achieving goals for blood pressure, central venous pressure, and central venous oxygen saturation. This includes getting the patient to a place where these therapies can be continued in a prompt manner. Most hospitals are able to respond in this early fashion for heart attack, stroke, and trauma patients. I think the same should be done for sepsis. We should treat sepsis the same way we treat these other conditions, because sepsis is a disease that affects just as many patients as these other conditions and sepsis patients are just as sick, if not sicker.

Medscape: Dr. Rivers, would you like to offer any parting thoughts?

Dr. Rivers: The take-away message is to understand that this is also preventive medicine. This is preventing progression of disease to the point where it is too complex to be treated. If you identify and treat patients earlier, you prevent the long-term consequences. Patients don't go on ventilators; they don't go on hemodialysis.[9] In addition, these patients don't go to long-term care facilities because of disabilities from their septic episode. I think, in the big picture, there is early detection, there is rapid intervention, and, again, there is a decreased incidence of sudden cardiopulmonary complications.

EGDT or the RB was never intended to be high-level critical care or doing procedures that are beyond the reach of a primary care physician or clinician. It is about early detection and rapid intervention which prevents downstream complications. Remember, the concept of early management is no different from acute myocardial infarction, stroke, or trauma.