Protocols can improve care, but the evidence behind the guidelines can be shaky by Deborah Gesensway
Published in the September 2007 issue of Today's Hospitalist
A study published this year in Critical Care Medicine put a number on visits from patients with suspected severe sepsis in U.S. emergency rooms every year: more than 500,000.
“Sepsis is bad and common,” pointed out Michael D. Howell, MD, a critical care physician who spoke at the Spring 2007 Hospitalist CME Series meeting in Redondo Beach, Calif. “But treating sepsis is four times more effective than revascularizing acute MI, in terms of number of lives saved.”
To help hospitalists save those lives, Dr. Howell, who is associate director of medical critical care at
“Treating sepsis is four times more effective than revascularizing acute MI, in terms number of lives saved."
—Michael D. Howell, MD Beth Israel Deaconess Medical Center
Beth Israel Deaconess Medical Center in Boston, outlined the commonly accepted cornerstones of effective sepsis management. But because controversy continues to swirl around many of those best practices, he said that physicians must navigate guidelines carefully.
That’s why in addition to discussing strategies that hospitalists can use to implement key management recommendations, he highlighted the latest findings, many of which call those same recommendations into question.
Making the call
Because sepsis can be so lethal, physicians need to be able to quickly determine which hospitalized patients are likely to develop serious sepsis. Not surprisingly, then, Dr. Howell said the first cornerstone of effective sepsis management is early identification and risk stratification.
Researchers are now developing tools to make the task easier. Dr. Howell described one prediction system, developed by a team at Beth Israel Deaconess, that emergency rooms can use to identify which patients are most likely to develop serious life-threatening sepsis during their hospital stay.
The Mortality in Emergency Department Sepsis (MEDS) score, described in the January 2007 issue of Critical Care Medicine, predicts 28-day and one-year mortality for emergency room patients with a clinically suspected infection. (See “MEDS score: a look at the parameters,” right.)
According to Dr. Howell, a MEDS score can predict illness severity much quicker than APACHE. Another plus: It uses data that are readily available in most emergency departments.
“Unlike the APACHE score, which requires 24 hours of data and other risk stratification scores, the only lab data here are an initial CBC for platelets,” Dr. Howell explained. “In most centers, those would be available in 30 to 60 minutes after showing up at the ER.”
However, it may be even easier to predict which patients need the most intensive care. Dr. Howell described another study from the same Beth Israel Deaconess group that focused on measuring hyperlactatemia alone as a predictor of mortality among septic shock patients.
Several recommendations from well-respected groups are based on those results. Guidance from the Surviving Sepsis Campaign and the Institute for Healthcare Improvement, for instance, state that “obtaining serum lactate is essential to identifying tissue hypoperfusion in patients who are not
yet hypotensive but who are at risk for septic shock.”
According to Dr. Howell, who is on the guideline committee of the Surviving Sepsis Campaign, experts recommend focusing on all patients with serum lactate levels greater than 4 mmol/L (36 mg/dL), regardless of blood pressure. (More information about serum lactate is online.)
“In my experience, anything that requires math at the bedside fails,” he said, even if that math is only simple addition as in the MEDS score. Measuring lactate avoids math altogether, framing the issue with simple questions: Is the blood pressure less than 90, or is the lactate greater than 40?
“If either of these is true,” Dr. Howell said, “you’re looking at a mortality of 28%. If these are false, you’re looking at a mortality of 2.5%. That’s pretty good for a 15-minute blood test and a sphygmomanometry measurement.”
The right antibiotic
Obviously, Dr. Howell said, treatment in the form of antibiotics is key. Unfortunately, using the wrong antibiotics may actually increase patient mortality.
He pointed to a study in the February 1999 issue of Chest in which ICU patients with sepsis experienced a four-fold increased risk of dying if they initially received the wrong antibiotics, as judged by subsequent cultures.
The Surviving Sepsis Campaign guidelines recommend appropriate diagnostic studies before starting antibiotics, along with the early administration of broad-spectrum antibiotic therapy. That timeline calls for therapy within the first hour of recognizing severe sepsis, but after obtaining appropriate cultures.
The guidelines then call for reassessing therapy with microbiology and clinical data to narrow coverage, when appropriate, and a typical therapeutic course of antibiotics lasting seven to 10 days. Physicians are urged to choose antibiotics according to local susceptibility patterns, a point that Dr. Howell said is critical.
“Early antibiotics and source control are keys to survival,” he concluded, noting that antibiotic choices have to be individualized. “But you want to at least consider MRSA and resistant gram-negative rods. At many centers, these are common and important pathogens even in patients coming from the community.”
Protocols in flux
Dr. Howell noted that several studies, including one from his own research group, have shown that the use of protocols in sepsis care results in earlier delivery of antibiotics, better antibiotic choices and faster resuscitation. When deciding on treatments, however, Dr. Howell warned hospitalists that much of the advice in protocols is based more on theory than evidence.
Although some treatments, such as hydration, vasopressors, antibiotics and mechanical ventilation, have been around as long as 80 years, many others have failed to withstand the test of time. Since the 1970s, for instance, he said, more than 50 trials of novel therapeutic agents have been tried for sepsis, and all have failed.
The question of what evidence is behind the most accepted protocols out there certainly applies to another management cornerstone: early goal-directed therapy. The study in the Nov. 8, 2001, New England Journal of Medicine that forms the basis of the recommendation in most sepsis protocols, including the Surviving Sepsis Campaign, prompts clinicians to begin immediate resuscitation in the emergency room, even before ICU admission. That appears to dramatically reduce in-hospital and 28-day mortality among sepsis patients.
“But should we change practice based on an N of 233?” Dr. Howell asked, noting the number of patients in the single-center randomized, controlled trial on which the guideline is based. As a result, he said, physicians need to wait for the results of a recently approved, NIH-funded multi-center research trial of the practice, which is now underway.
“Right now, the best existing data definitely say you ought to resuscitate people using early goal-directed therapy” Dr. Howell explained. “But recognize that a future multi-center validation may not bear that out.”
Saline or albumin?
In terms of which fluids are optimal, studies don’t offer much help. That’s because the literature favors both saline and albumin.
The best trial, which was done in Australia and New Zealand and published in the May 27, 2004, New England Journal of Medicine, seems to show that “if you received saline or if you received albumin, it didn’t matter for your survival.” Because saline is so much less expensive, he explained, it’s his preference.
And the wet vs. dry debate about how to manage fluids in sepsis is currently coming down on the side of the conservative strategy, thanks to an article in the June 15, 2006, New England Journal of Medicine.
Researchers found that patients with an acute lung injury who were run dry after the need for initial vasopressors had abated (a conservative fluid strategy) spent less time on the ventilator and fewer days outside the ICU than those who were run wet (a liberal fluid management strategy). The study found no difference in rates of either renal failure or mortality between the two groups.
Dr. Howell said that he recommends aggressively resuscitating patients who exhibit hypotension after a 1-1.5L fluid challenge or who have a lactate greater than 4 mmol/L plus an infection. He follows the early goal-directed therapy guidelines fairly strictly, setting a target of a central venous pressure of at least 8-12 mmHg, a mean arterial pressure greater than 65 and central venous saturation greater than 70%, and usually uses a crystalloid.
Once patients are resuscitated, however, he said to consider running them dry if they are still on the ventilator.
Controversies over activated protein C and steroids
Controversy continues to dog two other cornerstones of current therapy: activated protein C and steroids. The Surviving Sepsis Campaign guidelines recommend recombinant human activated protein C (rhAPC) for patients at high risk of death with an APACHE score of at least 25, multiple organ failure, septic shock or ARDS.
But studies offer contradictory advice on whether drotrecogin alfa (Xigris) helps only the sickest patients (with APACHE scores over 25) or patients who are not as sick—or no one at all.
According to Dr. Howell, he comes down on the side of at least considering drotrecogin for patients with a clear life-threatening infection and whose APACHE scores—personally calculated—exceed 24. “I never give it to anyone with single-organ failure, an APACHE of less than 24 or substantial risk factors for bleeding, although others do,” he said. “It remains to be seen what the official recommendations are going to be.”
When it comes to courting controversy, however, the clear winner goes to the Surviving Sepsis Campaign’s recommendations on using steroids.
The groundbreaking Annane trial, published in the Aug. 21, 2002, Journal of the American Medical Association, led to the recommendation that all patients in septic shock should receive steroids in relatively low doses, along with an ACTH stimulation test.
The problem, Dr. Howell said, is that experts are waiting for the results of a newly completed, much bigger and better designed study. An early abstract from that study seemed to contradict the 2002 study results, leaving the question of steroid therapy for sepsis very much up in the air.
A major problem with the 2002 study, Dr. Howell explained, was that the patients included were “not your run-of-the-mill patient.” Instead, they were very sick patients in vasopressor-resistant shock.
For these types of very sick patients who meet the criteria laid out in that trial, low-dose hydrocortisone (50 mg, four times a day) “does tend to reverse shock.” He added that patients who chronically take steroids or have other risk factors for adrenal insufficiency should be given them as well if they go into septic shock.
But other than these groups, Dr. Howell added, steroids probably shouldn’t be given, and an ACTH stimulation test seems to be unnecessary based on the newer studies.
Putting it all together
Less attention-grabbing, but “probably more important than all the other cornerstones combined,” said Dr. Howell, is the need to provide meticulous ICU care, including hand hygiene, the prevention of central-line infections, and elevating the head of the bed, among other interventions.
He described how his group at Beth Israel Deaconess has put all this evidence together into one pathway, called the Multiple Urgent Sepsis Therapies (MUST) protocol. The protocol is designed to be easily followed by all physicians and nurses involved in the care of these complex patients. The goal is to make sure that key aspects of state-of-the-art sepsis care aren’t forgotten in the rush of the moment.
The protocol uses flow sheets and is nurse-driven, so that nurses can give septic patients fluids without always having to query a physician. The protocol also ensures that appropriate antibiotics are given more quickly and that patients get better glycemic control.
A study of the effectiveness of that protocol, published in the April 2006 issue of Critical Care Medicine, found that fluid resuscitation, antibiotics and glycemic control all improved, while mortality fell from 29% to 20%. Although that drop was not statistically significant, said Dr. Howell, it was “at least headed in the right direction” and showed a very similar relative risk reduction to that in the early goal-directed therapy trial.
Finally, Dr. Howell told hospitalists to be prepared to change their protocols related to sepsis management. The Surviving Sepsis Campaign has just completed a major revision of its consensus guidelines, he pointed out, and new guidelines should be out in a matter of months.
Deborah Gesensway is a freelance writer reporting on U.S. healthcare from Toronto, Canada.
Check out the Today’s Hospitalist Clinical Protocol Web page for the protocol used for sepsis management at Boston’s Beth Israel Deaconess Medical Center.