Recent Advances in the Management of the Patient with Severe Sepsis

Social and Economic Implications of Severe Sepsis

“Doctors are relatively comfortable thinking about this notion of infection and systemic inflammatory response with severe sepsis at the intersection,” said Derek Angus, MD, MPH, of the University of Pittsburgh School of Medicine. “But when asked questions about how many patients actually have severe sepsis, it is difficult to find good answers.”

The Centers for Disease Control and Prevention have reported the incidence of septicemia annually since the 1970s. Over the last ten years, there were about 500,000 cases cited in those annual reports.

Publication of definitions during the 1990s resulted in studies estimating how many patients are diagnosed with severe sepsis. The first was an Italian study of 1,100 patients that found an incidence of about 5 cases per 100 ICU admissions (Salvo I, et al. Intensive Care Med. 1995;21:S244-S249). A later, unpublished paper by the same group revised that upward to 10 cases per 100 admissions.

Another by Rangel-Frausto screened all patients in three ICUs and three hospital floors for systemic inflammatory response syndrome (SIRS). They found one in five patients with SIRS progressed to severe sepsis. This is the first paper to look for severe sepsis outside the ICU, where they found about half of their cases (Rangel-Frausto, et al. JAMA, 1995;273:117-123).

“In looking at the literature, about 2 cases per 100 hospital admissions and 10 per 100 ICU admissions has been a consistent finding,” said Dr. Angus. “There is a lot of variation both within and across hospitals.”

To answer questions about this variation, Dr. Angus and others studied hospital admissions in seven states. According to their findings, there were about three cases per 1,000 of population or about two cases per 100 admissions. This is comparable to what the other studies reported. They also found about half of the cases outside of the intensive care unit (Angus DC, et al. Crit Care Med.2001;29:1303-1310).

“These numbers give you some sense of the magnitude of severe sepsis,” said Dr. Angus. “In 1995 there were 211,000 deaths from myocardial infarction. In that same year we projected about 215,000 died with severe sepsis. The new therapies, such as Xigris, may have an impact on mortality on the same scale as thrombolytics in myocardial infarction.”

This is only part of the story. What happens after the survivors leave the hospital is also of concern. There are important data suggesting septic patients appear to have late, ongoing mortality.

A study by Quartin (Quartin AA, JAMA.1997;277:1058-63) found only 20 percent of severe sepsis patients still living eight years later. Another from Perl found that 40% of severe sepsis patients leaving the hospital were dead within four years (Perl, et al. JAMA. 1995;274: 338-45)

Kaplan studied all patients enrolled in Medicare with a diagnosis of community-acquired pneumonia in the first quarter of 1997. They were followed for a year and then matched to hospital controls (Kaplan V. Ann Intern Med. 2002. In press).

“I found these numbers quite staggering,” said Dr. Angus. “One in three survivors of a hospitalization for pneumonia are dead at one year with odds of dying 2.5 times greater than the general population. Even at the twelfth month, they are dying at a rate almost 3 times the general population.”

The overall mortality curve shows rises in the elderly and a little increase in neonatal sepsis. There was also an increase in young males, largely related to HIV infections.

“On the other hand, looking at severe sepsis in those with no underlying disease, we see a much more smoothly rising curve,” he stated. “It is conceivable that a therapy targeted at in-hospital mortality may have no effect on long-term morality.”

In all cases and those managed aggressively, the most likely day to die is day one. It seemed to Dr. Angus that patients were arriving too late for effective treatment. Currently risk stratification is based on the person already being sick. If earlier interventions are to be used, there is a need to find markers for severe sepsis.

While admitting that there are no clear clues in this area, he did point to some early studies being developed relating genetic variations to severe sepsis. One adoption study found a five-fold increased risk if the biological parent had died of the same infection. There are a large and growing number of association studies, many relating to the tumor necrosis factor (TNF) gene.

Severe sepsis is not an inexpensive disease. Figures show that the average spent per hospitalization in the United States is about $22,000 or $17 billion a year. About half comes from Medicare, reinforcing the idea that this is a disease of the elderly.

“If you do a little bit of triangulation, one certainly gets the picture that severe sepsis is not an orphan disease,” said Dr. Angus. “It is very common, very expensive and frequently lethal.”

Pathophysiology of Severe Sepsis and Its Relationship to Acute Lung Injury

There have been major advances in understanding the pathophysiology of severe sepsis and why organ system dysfunction develops. This has lead to the development of a number of potential methods to treat, and possibly even prevent, this disease.

“One of the first mechanisms in the cascade is release of endotoxins and other cell wall components from both gram-negative and gram-positive bacteria,” said Edward Abraham, MD, at the University of Colorado Health Sciences Center in Denver.

This triggers the release of a number of mediators. Both use pathways that include toll-like receptors (TLR). Gram-negative products interact with TLR-4 initiating a sequence leading to the activation of multiple transcriptional factors, the most important of which is necrosis factor-kappa-B. Gram-positive organisms activate a similar cascade, primarily through TLR-2, that also leads to NF-kappa-B activation.

When cells see bacterial products, NF-kappa-B moves to the cell nucleus from the cytoplasm and binds to promoter regions of genes. Mediators known to be important in triggering severe sepsis are released in response to this movement.

“If we look at how this correlates with outcome, one sees that there are higher levels of NF-kappa-B activation in septic patients who die than in survivors,” said Dr. Abraham.

He looked at a group of patients with mild lung injury and another group on ventilators for 14 days or more or died. Those with a fatal outcome showed a greater activation of NF-kappa-B transcriptional factor. Those who survived showed much less activation.

Research has shown that agents against the endotoxin are not successful in clinical trials. While some of the reason may be the agents themselves, there are other possibilities.

“By the time we see these patients in the ICUs, the early wave of mediators had passed and we are treating them too late with these agents,” said Dr. Abraham. “We need to see the patients earlier.”

Another possibility is finding ways to interfere with the actions of events later in the cascade.

One possible target is activation of endothelial cells lining the blood vessels by pro-inflammatory cytokines. When triggered, they produce more cytokines and up regulate adhesion molecules to increase sticking of activated cells, such as neutrophils, to their surface. They also start the coagulation cascade.

“The role of this pathway in critically ill septicemia patients is still an open question,” stated Dr. Abraham. “We do know that this is a ubiquitous finding in critically ill septic patients.”

He cited the results of a study done several years ago. Although the agent being tested did not work, 75% of the patients had elevated prothrombin times and evidence of activation of the coagulation system when admitted to the ICU. In those patients the mortality was markedly increased.

Two other indicators of coagulation activation, increased D-dimer and decreased protein C levels, are found in almost all patients who present with severe sepsis. In addition, protein C levels appear to correlate with mortality.

“This would argue that replacement of protein C may be beneficial,” said Dr. Abraham. “The other issue that is raised is that correction of coagulation may be useful in and of itself.”

However, this does not appear to be the case. To illustrate, he noted results of two studies. The first was a double-blind placebo-controlled trial using high-dose antithrombin III in 2,314 adult patients with severe sepsis. There was no effect seen in 28-day all-cause mortality (Warren, BL, et al. JAMA 2001;286: 1869-1878).

Although the results were not published, a similar outcome occurred when giving patients tissue factor pathway inhibitor in the OPTIMIST trial. According to a November 2001 press release from the manufacturer, after approximately 2,000 patients were randomized to receive either placebo or tifacogin, there were no differences seen in the reduction of 28-day all-cause mortality and the trial was stopped.

“It would appear that the correction of coagulation alone doesn’t improve outcome,” said Dr. Abraham. “Since activated protein C does, its mechanism of action must be somewhat different.”

There are a number of other pathways involving activated protein C that may be important. It is an anticoagulant, increases fibrinolysis, and decreases neutrophil sticking. Perhaps most importantly, activator protein C’s interaction with the endothelial protein C receptor may lead to decreased endothelial activation.

“In addition to effects on coagulation, there are multiple other plausible and potentially important activities that this molecule [protein C] has in septic patients,” said Dr. Abraham.

He then turned his attention to two late-acting mediators that continue to rise hours and days after the initiation of an inflammatory response. One, started by cytokines, leads to generation of soluble phospholipase A2. This causes degradation of phospholipids, including surfactant, as well as activation of arcachidonic acid pathways.

Another other pathway of interest involves platelet-activating factor. This also degrades lipids, has endogenous inflammatory characteristics and can cause organ system dysfunction.

“If you give a healthy volunteer endotoxin, you see an increase in pro-inflammatory cytokines followed by activation of the coagulation cascade,” said Dr. Abraham. “There is a later rise in phospholipase A2 which may stay elevated for several days.”

Two recently described cytokine-like molecules that appear late are interesting to Dr. Abraham. Macrophage migration inhibitory factor (MIF) is released from multiple cells in the immune system and the brain. It activates macrophages and T-cells and appears to be a critical mediator of endotoxemia and gram-negative infection.

Since most patients with severe sepsis or septic shock present fairly late in their course, circulating levels of MIF are elevated in many of these patients and elevations correlate with outcome. In animal studies, using anti-MIF antibodies even at late points improves survival.

This suggests to Dr. Abraham that since MIF is elevated at the time patients are seen by physicians, anti-MIF antibodies may be attractive targets for
therapy.

The other mediator is high mobility group protein type 1. In mouse studies, release of this factor occurs 18 to 20 hours after endotoxin administration and circulates for days. Experimental models indicate that it can be blocked up to 24 hours after giving endotoxin and still salvage the animal.

“That’s an incredibly long period of time in these models and suggests that organ system dysfunction is still amenable to therapy after prolonged periods,” said Dr. Abraham. “It holds out the promise that we can combine agents working on multiple targets and get a synergistic effect.”

New and Investigational Interventions for the Patient with Severe Sepsis

The treatment of severe sepsis starts with using appropriate antibiotics, according to Robert Balk, MD of Rush Medical College.

Dr. Kollef and colleagues reported a significant increase in all-cause and infection-related mortality rates among critically ill patients given inappropriate antibiotic therapy at the start of the infectious process, irrespective of whether the right antibiotic was administered after the culture and sensitivity results were available. Successful management depends on knowing the bacterial ecology and antibiogram of your unit and starting the correct broad-spectrum antibiotic coverage (Chest 1999;114:462-74).

A recently published study from Rivers and others reinforced the importance of early goal-oriented resuscitation of the critically ill patient. They randomly assigned 263 severe sepsis and septic shock patients in the emergency department to an early goal-oriented treatment or conventional treatment strategy. The conventional arm targeted initial stabilization for transport to the floor or unit for care, while the early goal-oriented group had defined therapeutic targets for mean arterial pressure, central venous pressure, central venous oxygen saturation, and hemoglobin.

Using this early goal-oriented therapeutic strategy and not delaying the provision of necessary resuscitation and treatment resulted in a significant improvement in survival. Twice as many of the conventional therapy patients died of sudden cardiovascular collapse as compared to the goal-oriented therapy group (Rivers et al NEJM 2001;345: 1368-77).

Another interesting target for management of the critically ill patient involves the control of hyperglycemia. In a study of ventilated post-operative patients in Scandinavia, tight control of blood sugar (80-110 mg %) was associated with significantly improved ICU and hospital survival compared to patients allowed to have a blood glucose up to 180 mg %. There were four times as many deaths from multiple organ failure as a result of severe sepsis in the conventional glucose management group compared to the tight control patients (Van den Bergh et al NEJM 2001).

Attempts to interrupt or control the septic cascade with innovative treatment strategies over the past 20 years have enrolled close to 30,000 patients. A majority of these trials involved antimediator strategies, which have failed to demonstrate a survival benefit and some have been associated with the potential to cause harm.

The lack of success with antimediator therapy and the observation that the coagulation cascade is triggered in severe sepsis has prompted a change in therapeutic target. Fourier demonstrated decreased levels of antithrombin, protein C, and protein S in patients with severe septic shock. The nonsurvivors had lower levels of these natural anticoagulants compared to the survivors. (Fourier et al Chest 1993;104:882-88) This observation served as a basis for a number of recent investigations utilizing replacement doses of natural anticoagulants.

Recombinant human activated protein C, drotrecogin alfa (activated), was evaluated in the PROWESS (Recombinant Human Activa-ted Protein C Worldwide Evaluation in Severe Sepsis) Trial. This trial was stopped after 1,690 subjects were enrolled and a second interim analysis met the previously identified end-point for early stopping based on efficacy. The intention-to-treat primary endpoint, 28-day all-cause mortality, was reduced from 30.8% to 24.7%. This 6.1% absolute risk reduction was statistically significant. There was a relative risk reduction of 19.4% and the Kaplan Meier Survival Curves separated early and continued to separate.

This benefit was present over various demographic groups and with gram-negative, gram-positive, and mixed infections. Benefit was evident for the two most common sites of infection, pulmonary and intraabdominal. A greater effect was seen with greater numbers of failing organs. This was also evident with the high degree of illness as manifested by the third and fourth quartiles of APACHE (Acute Physiology And Chronic Health Evaluation) II scores.

In November 2001, The United States Food and Drug Administration approved the use of drotrecogin alfa (activated) for treatment of severe sepsis in patients with a high risk of mortality. To determine the optimum candidate with severe sepsis for treatment with activated protein C, Dr. Balk suggested that you ask the following four questions. If the answer to all is yes, the patient is a candidate for treatment.

1. Does the patient have severe sepsis and are they in the ICU?
2. Is there a commitment to provide full support?
3. Is the underlying disease process (other that the severe sepsis) treatable?
4. Is there an absence of risk factors for increased bleeding?