Despite maintaining adequate mean arterial pressure and central venous oxygen saturation, the mortality is still high in severe sepsis and septic shock. Previous studies have demonstrated that derangements in microvascular flow play a role in sepsis-induced multiple organ dysfunction and death. Lipopolysaccharide (LPS) or endotoxin is a specific ligand for Toll-like receptor 4 (TLR4), it can induce the following reactions including excessive immune and inflammatory responses , oxidative stress , capillary leakage, endothelial damage, impaired arteriolar and venular vasoregulation, and activation of the coagulation cascade 8. Subsequently, these reactions can lead to microcirculatory dysfunction. Polymyxin B adsorber hemoperfusion (PMX) have been proved to reduce mortality of severe sepsis and septic shock. Since 1994 to 2007, more than 60,000 patients have received this treatment. In a systematic review, the results show that PMX therapy was associated with significantly lower mortality risk (risk ratio, 0.53; 95% CI, 0.43 to 0.65). In a prospective, multicenter, randomized controlled trial (Early Use of Polymyxin B Hemoperfusion in Abdominal Sepsis [EUPHAS]), the results show that SOFA scores improved in the polymyxin B group, and 28-day mortality was 32% in the polymyxin B group and 53% in the conventional therapy group.

The investigators hypothesize that polymyxin B hemoperfusion can decrease blood endotoxin level and reduce endotoxin-related microcirculatory dysfunction. The purpose of this prospective, multicenter, randomized, controlled, open study is to investigate the effect of polymyxin B hemoperfusion on the sublingual microcirculation in patient with proven or suspected gram-negative bacteria severe sepsis and septic shock. The mean arterial pressure, dose of vasopressors and inotropics, SOFA score, PaO2/FiO2 ratio, and 28-day mortality will be investigated.

Introduction

Despite maintaining adequate mean arterial pressure and central venous oxygen saturation, the mortality is still high in severe sepsis and septic shock. Previous studies have demonstrated that derangements in microvascular flow play a role in sepsis-induced multiple organ dysfunction and death. Because of the clinical significance of microcirculatory dysfunction in severe sepsis and septic shock, recent advances in image technology have been used to investigate microcirculation. A sidestream dark-field (SDF) video microscope has been developed to visualize the microcirculation, but its measurement of microvascular blood flow classification is semi-quantitative. Previous studies have shown that a full-field laser perfusion imager can quantitatively measure the change of microcirculatory blood flow intensity. For clinical studies, sublingual microcirculation is most easily investigated using a SDF video microscope.

Because lipopolysaccharide (LPS) or endotoxin is a specific ligand for Toll-like receptor 4 (TLR4), it can induce the following reactions including excessive immune and inflammatory responses, oxidative stress , capillary leakage, endothelial damage, impaired arteriolar and venular vasoregulation, and activation of the coagulation cascade. Subsequently, these reactions can lead to microcirculatory dysfunction. Recently, a TLR4 antagonist, eritoran tetrasodium (E5564), has been reported to inhibit LPS response without TLR4 agonistic activity in animals as well as in humans, but it fails to reduce mortality of severe sepsis and septic shock. Polymyxin B adsorber hemoperfusion (PMX) have been proved to reduce mortality of severe sepsis and septic shock.Since 1994 to 2007, more than 60,000 patients have received this treatment. In a systematic review, the results show that PMX therapy was associated with significantly lower mortality risk (risk ratio, 0.53; 95% CI, 0.43 to 0.65).21 In a prospective, multicenter, randomized controlled trial (Early Use of Polymyxin B Hemoperfusion in Abdominal Sepsis [EUPHAS]), the results show that SOFA scores improved in the polymyxin B group but not in the conventional therapy group (change in SOFA, -3.4 vs -0.1; P =0.001), and 28-day mortality was 32% in the polymyxin B group and 53% in the conventional therapy group (unadjusted hazard ratio [HR], 0.43; 95% confidence interval [CI], 0.20-0.94; adjusted HR, 0.36; 95% CI, 0.16-0.80). Iba et al. have revealed that polymyxin B hemoperfusion can maintain better microcirculation and survival in a septic rat model.

The investigators hypothesize that polymyxin B hemoperfusion can decrease blood endotoxin level and reduce endotoxin-related microcirculatory dysfunction. The purpose of this prospective randomized controlled open study is to investigate the effect of polymyxin B hemoperfusion on the sublingual microcirculation in patient with clear or suspected severe sepsis and septic shock by gram-negative bacteria. The serum level of endotoxin, mean arterial pressure, dose of vasopressors and inotropics, SOFA score, PaO2/FiO2 ratio, and 28-day mortality will be investigated.

Methods

This protocol was reviewed by the National Taiwan University Hospital Research Ethic Committee(201208067RIB). Patients who meet the inclusion criteria and do not present any of the exclusion criteria will be randomly assigned to one of the two groups (the control group the PMX-HP group). In the control group, patients will receive standard treatment for sepsis according to the practice guidelines. In the PMX-HP group, patients will receive standard treatment for sepsis and polymyxin B hemoperfusion.

The hemoperfusion will be performed in the intensive care units within 24h after diagnosis of severe sepsis or septic shock. The 1st PMX-HP (day 0) will run for 2 hours. The 2nd PMX-HP must be performed within 24 to 36 hours after the end of the 1st treatment, ideally 24 hours.

The SDF video microscope (MicroScan, Microvision Medical) will be used to investigate total small vessel (less than 20 μm) density (TSVD), microvascular blood flow classification of each small vessel, perfused small vessel density (PSVD), microvascular flow index (MFI), and heterogeneity index (HI) of sublingual microcirculation. This device illuminates the tissues with polarized green light and measures the reflected light from the tissue surface. Both superficial capillaries and venules can be visualized because the scattered green light is absorbed by haemoglobin of red blood cells which are contained in these vessels. At each time point (0h, 24h, and 48h), five continuous image sequences (20 seconds) will be digitally stored, and data of the best three images (at least, if possible five) will be averaged for statistical calculation. The images will be analysed afterwards with automated analysis software (AVA 3.0, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands). Total small vessel density is automatically calculated by the software. A semi-quantitative method is used to classify the microvascular blood flow of each small vessel as follows: (0) absent (no flow or filled with microthrombosis), (1) intermittent flow (absence of flow for at least 50% of the time), (2) sluggish flow, and (3) continuous flow. (Figure 1) Small vessels with blood flow classification of (2) and (3) are considered as perfused small vessels, and the perfused small vessel density are automatically calculated. To calculate MFI score, the image is divided into four quadrants, and the same ordinal scale (0 to 3) is used to assess blood flow in each quadrant. The MFI score represents the averaged values of the four quadrants. The heterogeneity index is calculated as the highest MFI minus the lowest MFI divided by the mean MFI across the best three to five images at certain time point.

Safety

The blood pressure, body temperature, BUN, creatinine, ALT, AST, bilirubin (total/direct), lactate, sodium, potassium, and complete cell count will be monitored during the experiment. Adverse effects of polymyxin B hemoperfusion (bleeding, nephrotoxicity, neurotoxicity [agitation, weakness, drowsy, ataxia, numbness, blurred vision, and paresthesia], thrombocytopenia, and altered coagulation.) will be monitored.

Statistics

According to our previous study, the total small vessel density of the patients in the control group was 20.0 mm/mm2, 20 patients per group can detect a 12 % increase in total small vessel density of the patients in the PMX-HP group with an α level of 0.05 and β level of 0.8.

Data will be analysed using statistical software (SPSS 20; IBM SPSS, Chicago, IL). Means at certain time point between the two groups are compared using t test. Means from serial examinations between the two groups are compared using repeated measurement analysis of variance (with the factors of time and group) followed by Tukey or Dunnett's T3 multiple comparison tests. Data of MFI are expressed as median (interquartile range) and compared using the Mann-Whitney analysis following with adjustment for multiple comparisons. A P value < 0.05 is considered to indicate a significant result.

Inclusion Criteria:

Patients will be included in this study if they meet the following criteria (A+B+C):

A. Patients with the following conditions:

1. Abdominal cavity infection following emergency surgery.
2. Pneumonia, blood stream infection, urinary tract infection, or other infection, has received adequate treatment and presents with an Endotoxin Activity Assay > 0.6 EAA units.

B. SIRS, as defined by the presence of at least 2 of the following conditions (These criteria should have occurred between 12 hours before or 6 hours after the onset of the qualifying first organ dysfunction :

1. Fever or hypothermia (body temperature over 38 ℃ or under 36 ℃)
2. Tachycardia (heart rate > 90 bpm)
3. Tachypnea (respiratory rate over 20 breaths/min or under mechanical ventilation)
4. Leukocyte count more than 12,000 cells/mm3, less than 4,000 cells/mm3, or more than 10 % of immature form (band)

C. The presence of at least one of these symptoms of organ dysfunction or shock:

1. Cardiovascular system: an SBP of less than 90 mm Hg, a decrease in SBP of at least 40 mm Hg from baseline, a MAP of less than 65 mm Hg, or that requires treatments with vasoactive medication at any dosage.
2. Acute lung injury: PaO2 / FiO2 ratio less than 300 (ratio in mm Hg)
3. Acute kidney injury: creatinine more than 2 mg/dL, an increase in creatinine of more than 0.5 mg/dL, or diuresis of less than 0.5 mL/kg/h for 2 hours.
4. Acute liver injury: Total bilirubin level more than 4 mg/dL
5. Disseminated intravascular coagulation: platelet count less than 100,000 cells/mm3 or a reduction of more than 50 % of baseline
6. INR > 1.5 or aPTT > 60 sec
7. Altered mental status: GCS under 13 or 9T under endotracheal tube
8. Lactic acidosis: Lactate level more than 2 mmol/L (accompany with pH < 7.3 or Base excess < -5 mmol/L)

Exclusion Criteria:

Patients will be excluded if they

A. are under 20 years old or older than 99 years old

B. have suffered from severe sepsis or septic shock more than 24 hours

C. are pregnant

D. were treated with another medicine or device in the trial less than 30 days prior to the admission to this trial

E. have received organ transplantation less than 1 years prior to this trial

F. patients with hemophilia

G. have a allergic history of polymyxin B, heparin, or extracorporeal circulation

H. are terminally ill, for examples with metastasis, with a life expectancy of less than 30 days (certified by the attending physician)

I. have been diagnosed with HIV

J. present uncontrolled bleeding in the last 24 hours

K. were diagnosed with leukocytopenia (leukocyte count less than 500 cell/mm3) and/or thrombocytopenia (platelet count less than 50,000 cells/mm3)

L. have already received other blood cleaning treatments, such as CVVH, HD, HF, and PE upon entry into the trial

M. have a prior history of severe chronic organ failure

1. chronic respiratory failure ( COPD at last stage)
2. chronic heart failure (NYHA score = IV)
3. brain death
4. Chronic liver failure (Child Pugh score: C)

N. have evidence of infection by gram-positive bacteria, fungal infection, or mixed infection

O. have chosen palliative care and signed Do Not Resuscitate sheet

P. an APACHE II score over 30 present on entry into the trial

Q. non-native speakers