March 2011: Volume 37, Number 3
The State of Sepsis Care
Will Gains from Guideline Compliance Trump Procalcitonin?
By Genna Rollins
Even as researchers close in on a deeper understanding of the causes of sepsis and identify improved treatments and more specific biomarkers for the condition, experts emphasize that considerable ground can be gained simply by better implementing existing sepsis care guidelines, such as those of the Surviving Sepsis Campaign (SSC), an international initiative to improve the diagnosis, management, and treatment of sepsis. In particular, complete and consistent compliance with SSC’s sepsis care bundles— clusters of interventions that, when performed together, lead to better outcomes than if carried out separately—have the potential to make real inroads in reducing sepsis-related mortality. As integral members of hospitals’ sepsis response teams, labs have an opportunity to shine in these efforts.
Delivering on improved sepsis outcomes requires tight coordination by healthcare teams, according to Derek Angus, MD, MPH, professor and chair of critical care medicine at the University of Pittsburgh. “Hospitals often can do a good job complying with any individual bundle component, but overall bundle compliance is still less than ideal. That highlights how challenging it is to consistently deliver all the elements that we’d agree from an expert opinion level represent good care in the early management of septic shock and sepsis,” he said. “We’re looking at a team running through time-based instructions. Because there are key points where the team can’t initiate action until it knows results from the lab, the lab has to be included as an essential partner.”
An Urgent Effort
Sepsis, severe sepsis, and septic shock, a continuum of acute organ dysfunction associated with infection and the body’s inflammatory response, constitute the leading cause of death in non-cardiac intensive care units, with 30-day mortality rates between 28–50%. That grim statistic led to creation of SSC in 2002 with the goal of reducing sepsis-related mortality by 25% during the next 5 years. A collaboration of the European Society of Intensive Care Medicine, the International Sepsis Forum, and the Society of Critical Care Medicine, SSC has implemented a series of initiatives to reduce mortality and improve other outcomes associated with sepsis, including the first comprehensive clinical guidelines on sepsis in 2004, subsequently updated in 2008. SSC also partnered with the Institute for Healthcare Improvement (IHI) to develop a sepsis care quality improvement program. One of the most important deliverables from this effort involved organizing key components of the guidelines into bundles of care, a five-component resuscitation bundle and four-component management bundle (See Table, below).
Surviving Sepsis Campaign Care Bundles
The Surviving Sepsis Campaign (SSC) partnered with the Institute for Healthcare Improvement to develop sepsis care bundles based on key elements of SSC’s evidence-based sepsis guidelines. Many organizations have adopted these bundles in their own sepsis-related protocols.
Sepsis Resuscitation Bundle
- Serum lactate measured
- Blood cultures obtained prior to antibiotic administration
- From the time of presentation, broad-spectrum antibiotics administered within 3 hours for ED admissions and 1 hour for non-ED ICU admissions
- In the event of hypotension and/or lactate
>4 mmol/L (36 mg/dl):
— Deliver an initial minimum of 20 ml/kg of crystalloid (or colloid equivalent)
— Apply vasopressors for hypotension not responding to initial fluid resuscitation to maintain mean arterial pressure (MAP) >65 mm Hg
- In the event of persistent hypotension despite fluid resuscitation (septic shock) and/or lactate >4 mmol/L (36 mg/dl):
— Achieve central venous pressure (CVP)
of > 8 mm Hg
— Achieve central venous oxygen saturation (ScvO2) of >70%
Sepsis Management Bundle
- Low-dose steroids administered for septic shock in accordance with a standardized ICU policy
- Drotrecogin alfa (activated) administered in accordance with a standardized ICU policy
- Glucose control maintained > lower limit of normal, but < 180 mg/dl (10 mmol/L)
- Inspiratory plateau pressures maintained
<30 cm H2O for mechanically ventilated patients
Source: Institute for Healthcare Improvement website, accessed February 15, 2011.
Although it’s unclear if the effects are from increased awareness of sepsis or compliance with some or all of the SSC-IHI bundle elements, research shows an association between implementation of the bundles and improved sepsis outcomes, including mortality reductions. An analysis of data from 15,775 patients at 252 participating SSC sites found that unadjusted hospital mortality decreased from 37% to 30.8% during a 2-year period (Intensive Care Med 2010; 36:222–231). Individual organizations have reported improvements, too. For instance, bundle compliance during a 33-month study period at the Mayo Clinic in Rochester, Minn. rose from 12.7% at baseline to 52.7%, while mortality declined from 30.3% to 22% (Crit Care Med 2011; 39:252–58).
Although the SSC-IHI sepsis bundles have been well-publicized and adopted by many organizations, others have yet to do so. The first three elements in the resuscitation bundle—determining serum lactate values, obtaining blood cultures prior to antibiotic administration, and starting broad-spectrum antibiotics within 3 hours—depend on prompt action from the lab. Routine lab values, such as complete blood count, neutrophil count, and C-reactive protein, among others, also factor into sepsis-related clinical decision-making, even though they’re not officially included in the bundles.
For this reason, lab representatives should be at the table when protocols are being developed, according to Brad Karon, MD, PhD, director of hospital clinical labs at the Mayo Clinic. “Lab’s role is to work with clinicians to figure out what tests are needed at what point in time and help develop algorithms that pull out these things,” he explained. “The goal is to make it easy to order the right tests and get them back in a timely manner.” Karon was a faculty member for AACC’s January 12 webinar, “Sepsis and the Clinical Lab: Taking a Team Approach for Improved Outcomes.”
Organizations that don’t include lab directors in the protocol development process do so at their own peril, according to Gita Patel, RPh, PharmD, division director of clinical pharmacy services for HCA Supply Chain Services in Dallas. “We formed a multidisciplinary team to develop an algorithm and order sets. As we did so, we realized lactate kept popping up, and it was then that we contacted our lab. We’d assumed that whatever we’d need, they’d have right there. But we learned that our lab was sending this test out, so we couldn’t move further until this was rectified,” she recalled. “Learn from our lesson and include lab upfront.”
Community Hospitals Can Do It Too
Patel and her colleagues at The Medical Center of Plano (Texas), an HCA facility, compared outcomes for sepsis patients treated with and without care bundles (Ann Pharmacother 2010;44:1733–8). The results were striking: mortality in the non-bundle group was 61.1% versus 20% in the bundle group, and lactate measurements were available in just 9.4% of non-bundle patients compared with 79.7% of bundle patients. “A lot of community facilities are intimidated by both the Surviving Sepsis Campaign algorithm and the data collection required, and they think it can’t be replicated,” said Patel. “But we did improve the care of septic patients by applying academic principals to a community hospital.”
Patel’s experience should serve as a guidepost for other community hospitals, according to Todd Allen, MD, director of research in the department of emergency medicine at Intermountain Healthcare in Salt Lake City. “Of all the people who come down with sepsis, more than half go to non-academic medical centers. It’s in community hospitals where this work is taking place,” he said.
Allen explained that although Intermountain Healthcare’s flagship hospital, Intermountain Medical Center, employs a sophisticated rules-based screening algorithm that analyses all clinical inputs to give a real-time probability of septic shock for each patient, this information system has not been implemented throughout Intermountain Healthcare. The system’s rural and community hospitals “use a paper-based protocol instead of an electronic system that kicks in, and they’re still collecting data by hand to demonstrate outcomes. It takes a lot of people effort, but our preliminary data show that their outcomes are as good as our tertiary care facilities,” he said. Allen also served as a faculty member for AACC’s webinar on sepsis.
POC or Lab-based Lactate?
Laboratorians working to develop or improve sepsis response protocols will want to look carefully at current processes and determine how well they meet the organization’s goals for sepsis care. One consideration is whether turnaround times for lactate results are speedy enough with existing methods. Many organizations, including the University of Pittsburgh Medical Center, are happy with lab-based platforms. “Sepsis is a new patient population for rapid turnaround, but deserves the same attention as the other complex and seriously ill patients we treat here, and our lab has been a perfect partner in recognizing that,” said Angus.
Other institutions, including the Mayo Clinic, have gone with point-of-care lactate testing. The SSC-IHI quality improvement initiative calls for the resuscitation bundle to be completed within 6 hours, but Mayo Clinic added an additional target of completing the first three elements within an hour. Initially, Karon thought a lab-based lactate assay would fit the bill. “Our STAT lab was getting results within about 25 minutes from the time the sample was drawn, which was between 30 to 45 minutes from when it was ordered,” he recalled. “However, unless a patient is obviously septic, you need the lactate values to help decide next steps. We found we could do a point-of-care lactate within about five minutes, allowing most other decisions to proceed quickly so we could consistently meet that one hour goal. If we had continued to perform lab-based lactates, we would have had a 30 to 45 minute delay in that first hour before clinicians had the last key piece of information to decide the next steps.”
Still other hospitals, Intermountain Medical Center among them, haven’t taken the POC lactate plunge, due in part to concerns about POC analytical performance. “We did an experiment a few years ago comparing point-of-care versus lab-based lactate and found that point-of-care at that point wasn’t that accurate,” Allen recalled. “We may have made some mistakes in our study process and the technology has come a long way since then, but we’ve not tried to overcome the suspicion we created around point-of-care lactate. As it is, we’re getting about half of our lactate results in under 25 minutes.”
Karon also compared POC and lab-based lactate tests and found that two whole-blood POC assays correlated well with the lab-based reference method at normal ranges, but had a negative upward bias starting at values of 5–6 mM (Am J Clin Pathol 2007;128:168–71). However, as this imprecision was beyond the 4 mM sepsis-related cutpoint that Mayo Clinic was using, the organization was able to implement a POC assay with confidence. “Because we want to make decisions around a cutoff of 4 millimoles, we felt the point-of-care test enabled us to do that fairly reliably,” he said.
Whether using lab-based or POC lactate assays, labs also need to look carefully at pre-analytical issues related to lactate measurement. In the AACC webinar, Karon argued that these factors trump analytical concerns. A major consideration is that lactate isn’t stable in whole blood and levels will rise steeply and quickly if samples are left at room temperature. “Samples really can’t sit around for any length of time before they’re measured. Even five minutes is probably too long,” he said. Two possible solutions are to place samples in an ice slurry immediately and transport them post-haste to the lab, or to use tubes with glycolytic inhibitors.
Of course, laboratorians are well-aware of this issue. However, if they don’t have responsibility over phlebotomists or other personnel who will be drawing and transporting samples from suspected sepsis patients, a serious educational effort will be the order of the day, advised Dwayne Breining, MD, vice chair of hospital laboratory services at North Shore-Long Island Jewish Health System in Manhasset, N.Y. Even though this sprawling system’s 14 hospitals and other care facilities employ a variety of approaches to collect and transport specimens, use of sepsis care bundles made it easier to ensure consistently acceptable samples for lactate analysis. “With a rigorous algorithmic approach, we didn’t have nearly the problem with that that I thought we would. Whenever you roll out something from lab that requires special transport or handling, you’re gearing up for a headache. But since we rolled this out in an organized fashion, everyone was pretty much on board,” he indicated.
Tourniquet time and use, fist clenching at the time of blood draw, and patient exercise prior to the blood draw also can affect lactate values. Taken together, these pre-analytical issues are so significant that Karon believes they may have made the difference in some organizations’ choices around lactate assays. “One of the reasons labs may have poor success with point-of-care lactate testing is that nurses or others besides lab phlebotomists will draw the sample and do the test, and they may have assumed the device doesn’t work well when it’s really the pre-analytic issues,” he said. Indeed, another factor in Mayo Clinic’s decision to use a POC lactate assay was that it has a cadre of specially-trained vascular access technicians who not only are available to draw blood promptly but also well-versed in lactate pre-analytical considerations. “If we had to do the same thing with 3,000 nurses trained to do lactate tests, I’m not sure we would have come to the same conclusion,” he conceded.
What About Procalcitonin?
Although procalcitonin (PCT) is receiving a lot of interest as an earlier, more specific marker of sepsis, it has not been incorporated in SSC-IHI guidelines and does not appear to have been widely implemented in clinical practice. Part of the hold-up is lack of strong evidence for PCT’s role in identifying and managing sepsis. Although studies have linked the analyte to risk of developing sepsis and worsening pneumonia, and as a marker for ceasing antibiotic therapy, the evidence has not been strong and convincing to many physicians. “The reason procalcitonin doesn’t have a clear role right now is that it’s mainly and most successfully used as a rule-out test, not a rule-in test,” explained Angus. “If you’re already suspicious of sepsis, procalcitonin isn’t going to make you initiate the guidelines more quickly. It’s only going to give you advice on whether you can withhold antibiotics. I think procalcitonin has a potentially useful role, but we could do with more trials of it.”
Neither Mayo Clinic nor Intermountain Healthcare use PCT in their sepsis care bundles, mainly due to concerns about its overall sensitivity and specificity in diagnosing sepsis. However, both Karon and Allen also indicated that turnaround times for the test—50–55 minutes at Mayo Clinic and about 3 hours at Intermountain Healthcare—are too slow right now to fit in with their sepsis care bundles.
It takes about 30 minutes to run a PCT assay at Stanford University Medical Center in Palo Alto, Calif., but the test still does not have a place in the institution’s sepsis protocols, according to James Faix, MD, associate professor of pathology. “I’ve spent a lot of time working with our Surviving Sepsis Campaign team and they’re very interested in using it, but aren’t doing so yet because of uncertainty about what cutoffs to use, etcetera,” he reported. The situation is about the same at North Shore University Hospital in Manhasset, N.Y., according to Loring Bjornson, PhD, chief of the clinical chemistry lab. “It’s available if an individual physician wants to order it, but it’s not been officially incorporated in our sepsis panel,” he said. “The turnaround time is about an hour, but it’s an expensive test and the question of its relevancy and efficacy has not been decided. The verdict’s still out.”
Hospitals and labs interested in bumping sepsis response up a notch will do well to pour over their data carefully to identify procedural bottlenecks and opportunities to wring-out more efficiency, according to Allen. “We would see ebbs and flows in compliance, and we’d go back to the drawing board, find people who were doing well, and try to understand exactly what they were doing,” he recalled. As an example, when Intermountain Healthcare first implemented sepsis bundles, it had less-than-ideal compliance with obtaining lactate measurements. After reviewing system-wide procedures, the sepsis response team rolled-out a policy enabling phlebotomists and ancillary staff in emergency departments to request a lactate when it seemed appropriate for the clinical situation. “That’s a unique approach, but it’s turned out to really benefit our ability to draw those lactates and get results back in a timely manner,” said Allen.
The healthcare system-as-learning-lab also has paid-off for North Shore LIJ Health System. To beef-up sepsis response across-the-board, the system is implementing a code sepsis protocol individually in each hospital. This enables any member of the clinical team to call a code sepsis, which launches a rapid and coordinated team response. “In a large system, things tend to happen faster at the lower levels. While this can be disorganized and disjointed, there’s a method to the madness in that as the implementations roll-out we can identify and incorporate best practices,” said Breining.
One of those best practices is being implemented at Forest Hills Hospital, which initiated a code sepsis early on in the process. “We learned to put the tubes for lactate in the same bag as the ones for blood culture, so when someone does initiate code sepsis, it’s one-stop shopping and everything’s right there. The person drawing the blood can just grab and go,” Breining explained.
Patel emphasized that robust data collection is not only essential to improving performance but also crucial even to launching sepsis care protocols. “You can never collect too much data. Even in community hospitals, there are all sorts of people who can help with that,” she said. “We presented data in every forum we could, and the difference in outcomes that we documented was how we got support to implement our protocols.”
In addition to procalcitonin, many other biomarkers are being explored as earlier indicators of sepsis or better predictors of prognosis. Research efforts also are focused on gaining a better understanding of the underlying disease mechanism for sepsis, which still is somewhat of a mystery. With that deeper understanding will come treatment advances and updated sepsis care bundles, Faix predicted. “The bundles are very successful, but they’re not as good as they could be because they’re not getting at the pathophysiology of sepsis. They’re supportive still, because we don’t know the pathophysiology.”
Angus’s lab is evaluating a panel of biomarkers thought to reflect the different processes that are distorted in sepsis. “It’s a big interest of ours to see if novel biomarkers improve the diagnosis or therapeutic decision-making in sepsis,” he said.
Allen believes that evidence from both prospective and retrospective studies now in progress will lead to changes in recommended protocols. “I think we’re going to find simpler bundles in the next couple of years, where maybe there’ll be fewer elements, but the focus of the bundles will be the same,” he said.
As the field advances and discoveries give way to new therapies and updated care bundles, one thing is for sure: labs will remain on the front lines of diagnosing and treating sepsis.
Dr. Angus has received NIH funding to evaluate elements of the surviving sepsis initial management strategy and served as a consultant to companies developing biomarkers for sepsis.