Rivers et al. NEJM. 2001;345(19):1368–1377. | Randomized controlled trial of 263 patients with sepsis or severe sepsis assigned to receive six hours of early goal-directed therapy (EGDT) versus standard of care in the emergency room prior to admission to the intensive care unit. In-hospital mortality was the primary outcome. | Hospital mortality was 30.5% for the EGDT group versus 46.5% for the control group (p = 0.009). | Single-center study and bundled elements included transfusion of red blood cells in a way that has been proven to be unnecessary and possibly harmful. The utility of the central venous pressure to guide fluid responsiveness has been proven unreliable. | The importance of early and aggressive volume resuscitation in septic patients is emphasized. The difference in the two groups is essentially related to the amount of volume they each received in the first six hours. It is not clear that using the CVP and the hematocrit are ideal ways to manage sepsis resuscitation as there is strong independent evidence against both of them. |
Ibrahim et al. CHEST. 2000;118:146–155. | Prospective cohort study of 142 patients with blood stream infections admitted to intensive care units comparing hospital survival and adequacy of antimicrobial therapy. | Hospital mortality rate of patients on inadequate antimicrobial treatment (61.9%) was greater than the hospital mortality (28.4%; CI 1.77–2.69; p < 0.0001) of the patients on adequate antimicrobial treatment. | Single-center study. Mixed group of medical and surgical patients may not be applicable to other types of intensive care unit patients. Antibiotic decisions based on physician judgement, which may differ from places where guidelines/protocols are in place and mandatory. | Appropriate antibiotics are crucial in caring for patients with bloodstream infections. Suggests using broad coverage intitially and allowing culture data to narrow later to limit inappropriate use, which can breed resistance and not be cost effective. |
Gaieski et al. Crit Care Med. 2010;38:1045–1053. | Propsective cohort study of 261 patients with severe sepsis or septic shock undergoing EGDT to study the association between time to antibiotic administration and survival. | Mortality was markedly better for patients who received appropriate antibiotics both less than one hour from triage (mortality 19.5% versus 33.2%) and less than one hour from qualification for EGDT to appropriate antibiotics (25.0% versus 38.5%). | Single-center study. May not be applicable to other institutions not using an ER-initiated EGDT algorithm. | Early appropriate antibiotics are essential for patients with severe sepsis or septic shock. This cohort was treated with EGDT, which controls for volume resuscitation, and the vasoactive medication aspects of sepsis. The results support this important intervention. |
The NICE-SUGAR Study Investigators. N Engl J Med. 2009;360:1283–1297. | Mulitcenter randomized controlled trial of 6104 medical and surgical intensive care unit (ICU) patients expected to require ICU care for three or more consecutive days. Patients were assigned to intensive glucose control with target of 81–108 mg/dL or a conventional-control target of < 180 mg. Primary outcome was death at 90 days. | Significant mortality day 90 seen in patients in the intensive-control group versus the conventional- control group (27.5% versus 24.9%, p = 0.02). Severe hypoglycemia found in 6.8% of the intensive-control group versus 0.5% in the conventional-control group (p < 0.0001). | Open-label design, imbalance between the study groups with respect to number of patients on glucocorticoid therapy. | No additional benefit in lowering blood glucose levels below the range of approximately 140–180 mg/dL in critically ill patients. |
Sprung et al. N Engl J Med. 2008;358:111–124. | Mulitcenter, randomized, double-blind, placebo-controlled trial of 499 patients with septic shock with study patients assigned to 50 mg intravenous hydrocortisone every six hours for five days and tapered over six days versus placebo. Primary outcome was death at 28 days. | At 28 days there was no significant difference in mortality between the two groups with mortality 34.3% in the steroid group versus 31.5% in the placebo group (p = 0.51). There was no significant difference in mortality between the patients in the two study groups who did not have a response to cosintropin or between those who had a response to cosintropin. | The use of etomidate in 26% of the patients, which can inhibit corticosteroid production. Study lacked power due to early stop in patient enrollment due to slow recruitment, termination of funding, and time expiration of the study drug. | Based on this study, general use of corticosteroids for septic shock is not recmomended. As well, corticotropin testing cannot be recommended to determine which patients should receive corticosteroid therapy. Steroids did lead to more rapid reversal of shock, but no change in mortality. Subgroup with refractory septic shock showed a mortality benefit from steroids, similar to Annane study. Patients on hydrocortisone also had more cases of superinfection and hyperglycemia. |
Annane et al. JAMA. 2002;288:862–871. | Multicenter, randomized, double-blind, placebo-controlled trial of 300 patients with septic shock with study patients receiving hycrocortisone 50 mg every six hours intravenously and once daily administration of 50 mcg of fludocortisone via nasogastric tube versus indiscernible placebos for total seven days. Primary outcome was 28-day survival in nonresponders to the short corticotropin test. | There were 229 nonresponders to the corticotropin test and 70 responders to the corticotropin test with equal numbers of study and placebo in each group. In nonresponders, there was a significant (p = .02) decrease in death in the corticosteroid group and withdrawal of vasopressor therapy within 28 days in the corticosteroid group (p =.001) compared to placebo. No significant differences seen between responders. | The patients studied were in refractory shock, defined as systolic blood pressure < 90 mm Hg after one hour despite fluid resuscitation and vasopressor therapy. Significant number of patients received etomidate (24%) prior to enrollment, which is known to cause transient adrenal insufficiency. | Significant reduction in risk of death in patients with refractory septic shock given seven-day treatment with low-dose hydrocortisone and fludrocortisone with relative adrenal insufficiency (nonresponders to corticotropin stimulation). |
Bernard et al. (PROWESS study group).*
N Eng J Med. 2001;344:699–709. | Multicenter randomized, double-blind, placebo-controlled trial of 1690 patients with severe sepsis assigned to received intravenous infusion of activated protein C (drotrecogin alfa activated) for a total of 96 hours versus placebo. Primary end point was mortality at 28 days after initiation of the infusion. | The mortality was 30.8% in the placebo group versus 24.7% in the treatment group. Absolute risk reduction of death was 6.1% (p = 0.005). Serious bleeding was high in the group treated with activated protein C (3.5% versus 2.0%, p = 0.06). | Patients with high risk of bleeding (eg, chronic liver disease, chronic kidney failure on dialysis, thrombocytopenia) were excluded and these are commonly encountered in patients with severe sepsis. Majority improvement seen in sickest patients, those with APACHE II (Acute Physiology and Chronic Health Evaluation) scores > 25. | Activated protein C should be administered to patients with severe sepsis who meet the inclusion criteria. Recommendation cannot be made for patients with less-severe sepsis. |
Abraham et al. (ADDRESS Study Group).* N Engl J Med. 2005;353:1332–1341. | Multicenter randomized, double-blind, placebo-controlled trial of 2640 patients with severe sepsis and low risk of death (defined by APACHE II score < 25 or single organ failure) assigned to receive intravenous infusion of activated protein C (drotrecogin alfa activated) versus placebo. Primary end point was death at 28 days after initiation of infusion. | No statistically significant reduction in the 28-day mortality rate with the use of activated protein C (17% in the placebo group versus 18.5% in the treated group, p = 0.34). Rate of serious bleeding was greater in the activated protein C group than in the placebo group (2.4% versus 1.2%, p = 0.02) during the infusion. | Enrollment was discontinued due to the low likelihood of meeting defined objective. | Activated protein C should not be given to patients without severe sepsis, APACHE II scores < 25. Subsequent studies have shown some economic validation of the more ill patients in PROWESS, but have also demonstrated increased bleeding in patients receiving it as well. Treatment must be individualized, due to both side effects and cost effectiveness. |