Abstract
Abstract 4729
Sepsis and its sequelae remain leading causes of death in critically ill patients. Early identification and subsequently prompt treatment of sepsis can have a major impact on the outcome of septic patients. White Blood Cell (WBC) count, neutrophils count or percentage and percentage of band and immature neutrophils have been included in the diagnostic criteria of sepsis d. Previous studies have shown that morphological changes of leukocytes during sepsis can be quantified using Volume, Conductivity and Scatter (VCS) technology by some hematology analyzers. DxH800 (Beckman Coulter, Fullerton, CA) has the potential to report, in addition to Volume and Conductivity, five extra-laser diffraction angles for each cellular event.
The aim of the study was to evaluate the diagnostic accuracy of VCS parameters of leukocytes in detecting sepsis in critically ill adult patients.
This prospective observational study involved all consecutive adult patients admitted to a 31-bed medico-surgical department of intensive care (ICU) for more than 4 hours in a 3.5-month period. Diagnosis of sepsis was based on standard definitions. Blood samples were run by DxH800. Complete blood count (CBC) and WBC differential, but not VCS parameters were reported to attending physicians.
Of a total of 722 patients admitted during the observation period 422 had WBC differential (and thus VCS parameters) determined at admission, of whom 125 had sepsis Septic and non-septic patients had a median age (upper and lower quartile) of 60 (51-71) and 58 (45-72) years and APACHE II scores of 19 (14-25) and 11 (7-16). VCS parameters had higher Areas Under the Receiver Operating Characteristic Curve (AUC) than WBC or percentage of neutrophils for diagnosis of sepsis at admission (Table 1.).
DxH800 VCS parameters. Results as median values (upper and lower quartile)
| VCS parameters . | Sepsis N=125 . | No Sepsis N=317 . | p value . | AUC . | 95%CI of AUC . |
|---|---|---|---|---|---|
| Mean Volume of Neutrophils (MNVNE) | 153 (146–163) | 144 (140–148) | <0.001 | 0.78 | 0.72–0.83 |
| SD of Volume of Neutrophils (SDVNE) | 20 (18–23) | 17 (16–18) | <0.001 | 0.81 | 0.77–0.86 |
| Mean Axial Light loss of Neutrophils (MNAL2NE) | 155 (150–160) | 148 (144–151) | <0.001 | 0.78 | 0.73–0.84 |
| SD of Axial Light loss of Neutrophils (SDAL2NE) | 13 (12–16) | 11 (10–12) | <0.001 | 0.80 | 0.75–0.84 |
| Mean Volume of Monocytes (MNVMO) | 179 (169–189) | 163 (159–167) | <0.001 | 0.88 | 0.84–0.91 |
| SD of Volume of Monocytes (SDVMO) | 24 (21–28) | 18 (16–20) | <0.001 | 0.88 | 0.84–0.91 |
| SD of Axial Light loss of Monocytes (SDAL2MO) | 16 (15–20) | 13 (12–14) | <0.001 | 0.85 | 0.81–0.89 |
| WBC, 103/ml | 10.9 (7.3–15.3) | 10.5 (7.8–14.2) | NS | 0.51 | 0.44–0.57 |
| Percent of Neutrophils, % | 85 (76–91) | 80 (68–88) | <0.001 | 0.61 | 0.55–0.67 |
| VCS parameters . | Sepsis N=125 . | No Sepsis N=317 . | p value . | AUC . | 95%CI of AUC . |
|---|---|---|---|---|---|
| Mean Volume of Neutrophils (MNVNE) | 153 (146–163) | 144 (140–148) | <0.001 | 0.78 | 0.72–0.83 |
| SD of Volume of Neutrophils (SDVNE) | 20 (18–23) | 17 (16–18) | <0.001 | 0.81 | 0.77–0.86 |
| Mean Axial Light loss of Neutrophils (MNAL2NE) | 155 (150–160) | 148 (144–151) | <0.001 | 0.78 | 0.73–0.84 |
| SD of Axial Light loss of Neutrophils (SDAL2NE) | 13 (12–16) | 11 (10–12) | <0.001 | 0.80 | 0.75–0.84 |
| Mean Volume of Monocytes (MNVMO) | 179 (169–189) | 163 (159–167) | <0.001 | 0.88 | 0.84–0.91 |
| SD of Volume of Monocytes (SDVMO) | 24 (21–28) | 18 (16–20) | <0.001 | 0.88 | 0.84–0.91 |
| SD of Axial Light loss of Monocytes (SDAL2MO) | 16 (15–20) | 13 (12–14) | <0.001 | 0.85 | 0.81–0.89 |
| WBC, 103/ml | 10.9 (7.3–15.3) | 10.5 (7.8–14.2) | NS | 0.51 | 0.44–0.57 |
| Percent of Neutrophils, % | 85 (76–91) | 80 (68–88) | <0.001 | 0.61 | 0.55–0.67 |
On ICU admission, MNVMO had the highest discriminant values in detecting sepsis with sensitivity (Se) of 84%, specificity (Sp) 76%, positive predictive value (PPV)58%, negative predictive value (NPV) 92% and odds ratio (OR) 17 for a cutoff of 168. MNVMO<168 (59% of patients) gave reasonable evidence against sepsis (likelihood ratio (LR) of 0.21, pre- and post-test probability, 0.28 and 0.076,) and MNVMO≥175 (25% of patients) gave reasonable evidence for sepsis (LR of 6.85, pre- and post-test probability, 0.28 and 0.73, respectively). Only 16% of patients, in the intermediate range, had a rather uninformative test result (LR of 1.32). Combining monocyte or neutrophil VCS parameters with CRP slightly increases AUC to 0.91.
During the ICU stay, VCS parameters were obtained in 316 initially non-septic patients, of whom 20 became septic. MNVMO < 174 excluded ICU-acquired infection with NPV of 99% and MNVMO > 174 predicted ICU-acquired infection 1.7 days earlier than clinical diagnosis (Se: 85%, Sp 68%, OR 12 and AUC=0.78, 95% CI: 66–90).
This large prospective study in unselected critically ill adult patients demonstrates the diagnostic utility of VCS parameters, especially MNVMO, in detecting and, more importantly, excluding sepsis at admission and during ICU stay. VCS parameters are obtained automatically, routinely, within minutes, requiring neither additional sampling nor additional cost to that of complete blood count thus making their prospects very promising.
Pradier:Beckman Coulter: Consultancy, Membership on an entity's Board of Directors or advisory committees.
Author notes
Asterisk with author names denotes non-ASH members.
