Beyond the Horizon: Applying Today’s Proven Therapies to Increase Survival in the Severe Sepsis Patient

This program was supported by an unrestricted educational grant from Eli Lilly and Company.

Faculty disclosures for this summary are provided on page 12 of this publication

The State of Therapy for Severe Sepsis: Current Controversies in ICU Medicine

Sepsis remains a major cause of hospital-based mortality and morbidity. Severe sepsis, defined as sepsis associated with acute organ dysfunction, develops in more than 750,000 U.S. patients a year with a 30-50% mortality rate.

“Sepsis is a very old word,” said Jean-Louis Vincent, MD, PhD, Professor of Intensive Care Medicine at the University of Brussels, Belgium. “In ancient Greek already it meant disintegration of the human body.”

The Sepsis Occurrence in the Acutely Ill Patient (SOAP) study enrolled 3,147 patients hospitalized in 198 centers in Europe. They found that 37.4% of Intensive Care Unit (ICU) patients have an infection at one time or another during their stay. Lungs were the most common sites of infection. A very wide variety of causative microorganisms were seen (Vincent JL, et al Am J Respir Crit Care Med. 2003; 167:A837 [abstract]).

“This is an oversimplification, but host response is what sepsis is about,” said Prof. Vincent. “Infection is a microbiological invasion of our body and sepsis is the battle that rages when the body responds to that aggression. The septic response may become excessive like when an army creates trouble in the country it is supposed to defend.”

One way to bring order to the battlefield would be adoption of the PIRO classification system. P represents the predisposition of patients for responding to infection in different ways, based on several factors including genetics and co-existing disease. I represents the different types of infection that give rise to a range of responses in individual patients. R represents response as seen by different markers of inflammation. Finally, O represents organ dysfunction.

“The PIRO system gives us a framework to order our thoughts and hopefully make these issues less confusing at the bedside,” said Prof. Vincent.

Until recently, the treatment for severe sepsis was primary infection control, antibiotics, hemodynamic stabilization with fluids, and vasoactive agent. New therapeutic modalities that modulate the septic response have brought about a change in strategy.

The Recombinant Human Activated Protein C Worldwide Evaluation in Severe Sepsis (PROWESS) Trial was a randomized, double-blind, placebo-controlled trial looking at rh-APC (drotrecogin alfa [activated]) in sepsis. There was a significant decrease in 28-day mortality when comparing drotrecogin alfa (activated) to placebo, with the divergence beginning after a few days. The significant differences in survival continued at three months and at 30 months (Bernard GR, et al. NEJM. 2001;344: 699).

When looking at the relative risk of death, there was greater benefit seen in patients with an APACHE-II score of 20 to 24. Prof. Vincent did note, however, that the confidence intervals for the lower scores crossed the zero line. It is not possible to say that the drug is ineffective in patients with milder forms of the disease (Ely EW, et al. Crit Care Med. 2003;31:12).

“When we looked at patient location over the 28-day period of study, patients treated with the drug did not stay longer in the ICU or the general ward,” stressed Prof. Vincent. “Patients treated with drotrecogin alfa (activated) were weaned off vasopressor agents earlier.” In addition, those on the medication had a more rapid resolution of their respiratory dysfunction (p=0.009) (Vincent JL, et al. Crit Care Med. 2003;31:834).

Hemorrhage is, of course, a concern. A recent report with complete worldwide experience with the drug in more than 2,700 patients indicated that the number of cerebrovascular accidents (CVA) remains small. Furthermore, the number of thrombotic events may be reduced, so that the global numbers of CVA may be lower, but the numbers in this part of the study are very small (Bernard G. Crit Care. 2003;7:155).

“Cancer may be a good analogy for sepsis,” said Prof. Vincent. “Both are dreadful diseases that can affect any area of the body and involve both surgical and medical therapies. Exactly as in the treatment of cancer, I think in the future that we will use combination therapies.

Cost-Effectiveness of Current Therapies

When determining the value of any therapeutic intervention, both costs and outcomes should be considered. Total expenditures for treating severe sepsis can be assessed in the context of quality-adjusted life years (QALY). QALYs provide an indication of the benefits gained from a medical procedure in terms of quality of life and survival for the patient.

“Cost-effectiveness analyses (CEA) help delineate and clarify difficult choices, but do not necessarily make the choices any less difficult to make,” said Derek C. Angus, MD, MPH, Professor and Vice Chair, Department of Critical Care Medicine, at the University of Pittsburgh School of Medicine.

Essentially cost-effective analysis is a ratio of the change in cost going from one care plan or intervention to another divided by the change in effects. The best outcome is where you save lives and money, but this rarely happens. One must decide if the money spent is worth the benefits gained. There are differing amounts of uncertainty that also have to be factored in.

“If the point estimate and most of the uncertainty appears to be better than $X per QALY and you have a health care system that can spend up to $X per QALY, then you would look at this hypothetical intervention and say we can afford it,” said Dr. Angus.

Cost-effectiveness analysis is only useful under the assumptions that the budget for health care spending is limited and that it is not sufficient to pay for all possible interventions in all possible scenarios. Given these assumptions, one can rank interventions based on their comparative cost-effectiveness and then allocate resources to the most cost-effective therapies first. This is a utilitarian or distributive justice model where one tries to spend money on those things that give the biggest return on investment first and then gradually go down the list until the money is gone. In theory, this results in the most efficient use of health care resources.

A therapy that is effective but associated with a high cost-effectiveness ratio (e.g. >$500,000/QALY) may still be acceptable to society and affordable if it applies to a small number of patients. Problems arise when it is to be used for a condition that affects a lot of people, such as sepsis or Adult Respiratory Distress Syndrome.

Even when implementing adoption of a cost-effective therapy, it can quickly become tricky. For instance, a new therapy in the ICU may be cost-effective but expensive. If, for example, one of the existing programs is a surgical procedure by a powerful surgeon, it might be problematic suggesting that the program be cancelled so funds can be used to support the new ICU drug.

Costs and outcomes also vary depending on one’s perspective. Society might be willing to spend some specific amount of money per QALY. The patient is willing to spend a different (probably much greater) amount. So, the cost and outcomes ratio would be different depending on the point of view.

“The responsibility of the physician is probably to adopt a patient or societal perspective,” said Dr. Angus. “Hopefully society and the patient will have shared views. The problem is that doctors, consciously or not, often adopt the point of the view of the institution to help it save money, etc. This is almost certainly inappropriate.”

When considering sepsis care, one element is provision of organ support. An example is an analysis of the cost-effectiveness of hemodialysis for acute renal failure (ARF) conducted by Mary Beth Hamel and colleagues. They found that patients who developed ARF in the SUPPORT study had 0% survival without dialysis and a 27% survival with dialysis; although 33% of the survivors became dialysis dependent. Despite the huge gain in survival, the initiation of dialysis cost $180,000 per QALY (expressed in 2003 US$), which is very expensive (Hamel MB, et al. Ann Intern Med. 1997; 127:195).

The reason for the poor cost-effectiveness ratio is the high set of entrained, downstream costs associated with the decision to initiate hemodialysis. “Most of these costs are completely invisible to the clinicians,” said Dr. Angus. “We initiate treatments without thinking about the downstream costs and consequences that are involved.”

Dr. Angus compared this analysis to recent studies of the cost-effectiveness of drotrecogin alfa (activated) for sepsis (Angus DC, et al. Crit Car Med. 2003; 31:1; Manns B, et al. NEJM. 2002; 347:993).

For both dialysis use in ARF and drotrecogin alfa (activated) use in severe sepsis, the cost-effectiveness is very sensitive to case selection. Overall, use of drotrecogin alfa (activated) appears more cost-effective than initiating dialysis. Yet because drotrecogin alfa (activated) use is associated with new or additional outlays, it is perceived as more expensive.

Ultimately, clinicians may be very uncomfortable initiating or withholding therapies at the bedside based solely on cost-effectiveness. If clinicians wish to avoid such thoughts, they must entrust the system to put in place some policy that ensures rational provision of services tailored to society and patient preferences.

“Such a policy is probably impossible without input from clinicians,” said Dr. Angus. “Consequently, clinicians have to engage in thinking about cost-effectiveness whether at the bedside or through policy. CEAs can help the clinician by clarifying costs and consequences as well as by providing a framework for viewing the tradeoffs of alternative decisions. It confronts the reader with the realities of resource allocation and serves as an ethical roadmap when societal or patient worth is at odds with the hospital or provider perspective.

Case-Based Learning: Illustrative Cases

Recent studies have shown the promise of significantly reducing mortality in severe sepsis patients. To illustrate the thought processes behind clinical decision-making in sepsis patients, Gregory A. Schmidt, MD, Professor of Medicine, Anesthesia and Critical Care, the University of Chicago, discussed two case studies.

The first was a 56-year-old woman with history of alcohol and cigarette abuse who presented with shortness of breath, cough, fever, sputum, and a typical history of pneumonia. Laboratory studies included 28% bands, modest coagulopathy, thrombocytopenia, acidosis, renal dysfunction, and elevated blood sugar.

Non-invasive ventilation was attempted, but she failed and was placed on ventilation. A chest x-ray following intubation revealed a large area of consolidation at the left base of the lung. A tap of the area to assess for a para-pneumonic perfusion produced no fluid. A follow-up x-ray showed a progression to bilateral infiltrates.

“The one intervention being considered at this point with very little controversy is use of low tidal volume,” said Dr. Schmidt. “A very well done
and important study by The Acute Respiratory Distress Syndrome Network showed a clear difference in mortality in the patients with lower tidal volumes when compared with more traditional volumes. We should be applying this regularly in our ICUs, but it appears as though we aren’t” (The Acute Respiratory Distress Syndrome Network. NEJM. 2000; 342:1301).

The patient met the standards for admission into the PROWESS trial, which was being used as criteria for considering the use of drotrecogin alfa (activated). However, given the history of alcohol abuse, low platelet count, and nasogastric aspirate positive for blood, is she a good candidate for the medication?

The bleeding risk from the PROWESS study was 3.5% in study group compared with 2% in placebo. The increased risk was seen only during the infusion period and mostly occurred when there was a clear underlying disposition or following surgery.

“Some look for reasons not to give new therapies, but I don’t agree with that response,” said Dr. Schmidt. “The bleeding was not a very big problem overall and not a reason to forego use of drotrecogin alfa (activated) in this patient.”

Over a period of time the septic shock resolved. There was improvement in renal dysfunction, she was weaned from vasoactive drugs by day four, extubated by day eight and was eventually discharged.

“She was the poster child for use of drotrecogin alfa (activated),” said Dr. Schmidt. “With the exception of the addressed coagulation concerns, this was a slam dunk.”

The other patient was a 49-year-old woman who had had a hysterectomy nine days previously. She was discharged to home with a surgical drain and given instructions to come to clinic for drain removal after five days.

She did not return and was found obtunded nine days after surgery by family members. On examination she was febrile, hypotensive, tachypneic, unresponsive with laboratory findings including bandemia, renal dysfunction, metabolic acidosis, elevated ammonia level, and hyperbilirubinemia. An abdominal film showed free intra-abdominal air.

After a round of antibiotics she was taken to surgery. The surgeons found a purulent hematoma at the vaginal cuff and infection throughout the abdomen.

“She had bleeding around the time of her first operation, a hematoma at the vaginal cuff, and was just coming from another major operation,” said Dr. Schmidt. “As you can imagine, there was a delicate discussion involving the intensive care physicians and the surgeons about whether to give a drug that might have a risk for bleeding.”

It was decided to start the drotrecogin alfa (activated) after the surgeons thought they achieved adequate homeostasis. The doctors involved saw little reason to not give a therapy that all thought could be helpful. Twelve hours after surgery, the medication was begun.

By the second hospital day, the patient had developed acute renal failure and dialysis was started.

“This complicated the issue of drotrecogin alfa (activated) therapy by necessitating an invasive procedure to get access,” said Dr. Schmidt. “However, we did not want to risk losing the benefits of the medication. In response, we stopped drotrecogin alfa (activated) for two hours while the dialysis catheter was placed.”

Dr. J.P. Kress and others undertook a study of sedative use in ventilated patients, many with sepsis. They would stop sedatives daily until the patient either became agitated or was able to follow simple commands. If the medications needed to be restarted because of agitation, the dose was halved each day (Kress JP, et al. NEJM. 2000;342:1).

“That simple intervention significantly reduced the duration of mechanical ventilation from seven to five days,” noted Dr. Schmidt. She was also successfully treated and released from the hospital.

“When trying to decide which therapies to incorporate in our practices, I usually look at how well the study was done, its biological plausibility, the depth of the supporting data, and existence of multiple studies,” said Dr. Schmidt. “Secondly, consider safety and perhaps take extra steps if needed. Finally, when there may be risks that are acceptable in a patient, involve the surgeons and others ahead of time so we don’t second-guess ourselves should something go wrong.

Faculty Disclosures:

Derek C. Angus, MD
Reported nothing to disclose

Gregory Schmidt, MD
Consultant: Eli Lilly
Speaker: Eli Lilly

Jean Louis Vincent, MD 
Honorarium: Eli Lilly, Chiron, Icon, Baxter