| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Abstracts of Oak Ridge Posters |
1
Department of Lab Medicine and Pathobiology and Department of Surgery, Toronto General Hospital, University Health Network, University of Toronto, Toronto, Ontario M5G 2C4, Canada
a address correspondence to this author at: Dept. of Clinical Biochemistry, ES-3-404, Toronto General Hospital, 200 Elizabeth St., Toronto, Ontario M5G 2C4, Canada
Endotoxin, or lipopolysaccharide (LPS), is the major gram-negative bacterial cell wall toxin that triggers septic shock (1). Gram-negative endotoxin is shed from the membrane of rapidly proliferating bacteria, and enhanced release into the blood is associated with antibiotic use (2). Endotoxin is also translocated through the gut after periods of hypotension often associated with cardiopulmonary bypass or hypovolemic shock (3). LPS has been demonstrated to enter the systemic circulation from the lung in experimental animal studies and human clinical investigations. In the blood, LPS binds to a carrier protein, which acts as a chaperone to carry this molecule to CD14 receptors on immunocompetent cells to trigger pro-inflammatory cytokine synthesis (tumor necrosis factor).
To date, the detection of LPS in blood or plasma has been largely dependent on variations of the limulus amebocyte lysate (LAL) assay, which utilizes the clotting enzyme cascade extracted from the primitive "white cell like" amebocyte cells of horseshoe crabs to detect LPS. The effectiveness of this assay in human blood and blood plasma has been controversial and problematic because of numerous interfering reactions and variations in between-lot and between-manufacturer reagent performance.
We have developed a rapid, homogeneous assay for the detection of endotoxin activity (EA) in whole blood based on in vitro neutrophil activation (4). This novel type of assay uses the priming effects of complement opsonized immune complexes on the respiratory burst activity of neutrophils as an analytical platform. Hypochlorous acid generated by the concerted activity of membrane-bound NADPH oxidase and azurophil granule myeloperoxidase of the neutrophil produces luminol chemiluminescence. Although immune complexes formed by the interaction of antibody with antigen do not directly stimulate neutrophil respiratory bursts, they prime neutrophil oxidative machinery to higher activity, which is subsequently elicited by challenging the cells with yeast cell wall zymosan.
The EA assay consists of three tubes containing lyophilized reagents that reconstituted with buffer (1 mL) containing luminol and zymosan. Assay reactions are initiated by the addition of 0.04 mL of whole blood to each tube, and light emission is monitored and integrated over 20 min in a temperature controlled (37 °C) photon counting luminometer (Berthold 953). One assay tube (blank) reflects baseline neutrophil activation in the absence of exogenous immune complexes. A second tube (test) contains a specific anti-LPS IgM that stimulates neutrophil activity in proportion to the concentration of LPS in the blood. The third tube (max) contains specific anti-LPS IgM and an excess of LPS so that the chemiluminescence reflects the maximum response of the individual patient sample. The differences in neutrophil activation and cell count between individual samples are normalized by subtracting the light integral of the blank from the test and max tubes and expressing the EA as the ratio of the test (minus blank) to the max (minus blank).
EA is a hyperbolic function of the LPS concentration over the range of 0–800 ng/L. The sensitivity of the assay, primarily dependent on antibody concentration, was adjusted to yield 50% of maximal signal in the steepest portion of the dose–response curve at borderline pathological endotoxin concentrations (50 ng/L). Detection of 20–30 ng/L of Escherichia coli 055:B5 LPS was routinely achieved in donor blood with added LPS, with signals >3 SD above the mean of parallel samples without added LPS. The assay demonstrated broad cross-reactivity with the LPS from pathogenic strains of E. coli, Pseudomonas, Salmonella, Klebsiella, and Serratia and was generally reactive with the widely conserved lipid A structure present in the LPS of most gram-negative organisms. The assay was not reactive with the cell wall products of gram-positive bacteria, including preparations of pathogenic strains of heat-killed Staphylococci and Streptococci, lipoteichoic acids, or fungal cell wall extracts.
Minimal blood constituents of the assay include blood plasma with an active classical and alternative pathway complement cascade and neutrophils or equivalent cultured phagocytic cells (retinoic acid differentiated HL-60 cells). Heat treatment of plasma to 60 °C for 10 min (which destroys complement activity) completely abolished antibody-dependent LPS signal detection.
Previous studies (4) have demonstrated that the assay is insensitive to variations in neutrophil count (0.5–20 x 109 cells/L) and erythrocyte concentration (0–140 g/L hemoglobin). Recovery studies using whole blood, from healthy volunteers, supplemented with increasing concentrations of LPS have demonstrated reproducible recovery using the EA assay but inconsistent recovery by acid extraction LAL methodology (4).
Assay specificity was confirmed in studies on patient blood samples obtained from the Medical-Surgical Intensive Care Unit at the Toronto General Hospital after informed consent and approval by the Hospital Ethics Committee. Nineteen blood samples that were endotoxemic by the chemiluminescent EA assay were reassayed after treatment with polymyxin B sulfate, a well-characterized endotoxin-binding agent. Pretreatment of the blood samples with 0.5 mg/L polymyxin B sulfate eliminated the endotoxin-dependent signal in 16 samples and diminished a previously high signal to low endotoxin concentrations in 3 samples. The assay was not affected by lipemia (<15 g/L triglycerides), hemolysis (<5 g/L hemoglobin), icterus (<200 mg/L), or variations in hematocrit from (20–60%).
Because of the inherent instability of blood samples in the assay
beyond 90 min at room temperature, pooled within-run precision was
evaluated by simultaneous assay of samples with and without added LPS
by three operators on three separate luminometers in replicates of
eight. A total of 10 samples (5 without added LPS, 5 with added LPS at
800 ng/L) were evaluated with a pooled CV of 12% achieved in samples
without added LPS and 8% in samples with added LPS. Within-run
precision was also evaluated in 200 intensive care unit (ICU) patient
samples assayed in duplicate at two independent clinical trial sites. A
within-run precision profile of CV vs EA indicated that CVs 
10% were
achieved across the dynamic range in >75% of samples, with a pooled
overall CV of 9.3% at both sites.
In an initial clinical trial conducted at three academic centers, a cohort of 169 ICU patients was evaluated. After approval by the hospital ethics committee and informed consent, patients with suspicion of infection were enrolled into the trial, and EA was assayed daily for a maximum of 7 days by the chemiluminescent assay. The first endotoxin assay result was determined within 8 h of ICU admission and compared with microbiological culture from any body site, including blood harvested within a 12-h window before the endotoxin assay. The following assay performance characteristics were achieved: clinical sensitivity for detection of gram-negative infection in any body site (sputum, blood, abdominal, thoracic or pelvic abscess drainage, urine, bronchoalveolar lavage, tissue biopsy) was 96% (47 of 49), the negative predictive value was 96% (49 of 51), the specificity was 43% (49 of 114), and the positive predictive value with regard to a gram-negative infection was 45% (53 of 118). The seemingly low positive predictive value and specificity was attributable to the frequent presence of endotoxin in the blood in the absence of a culturable gram-negative organism. This phenomenon can occur in clinical situations where prior antibiotic use stuns or kills microorganisms, preventing their growth in culture media, or in situations where translocation of LPS through leaky gut or lung epithelium can occur. Persistent endotoxemia in these situations is likely to have major clinical relevance and is the subject of on-going clinical trials.
In a subset of 74 patients from the clinical trial above, LAL assays
were also included, using the whole blood method of Tamura et al.
(5). No correlation between the LAL result and the
chemiluminescent endotoxin assay was evident. The chemiluminescent
endotoxin assay was significantly higher in patients with gram-negative
infection and sepsis syndrome, whereas the LAL assay was not able to
discriminate these patient subgroups (Table 1
). Failure of the LAL assay to correlate with gram-negative
bacteremia in patients with sepsis has recently been documented
(6).
|
We conclude that this novel method for the analysis of EA in whole blood exemplifies a generic analytical platform that incorporates the specificity of antibodies with the responsiveness of a patient’s immunological effector cells. Within minutes, this EA assay yields key information about the infective status of the patient that may be of prognostic and diagnostic value and may not be otherwise available by microbiological culture for several days.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
