WHITE BLOOD-CELL COUNTS AND PLASMA C3A HAVE SYNERGISTIC PREDICTIVE VALUE IN PATIENTS AT RISK FOR ACUTE RESPIRATORY-DISTRESS SYNDROME

Objective: To investigate and select nonassociated variables with pred ictive Value for acute respiratory distress syndrome (ARDS) in patient s at risk, Design: Prospective, observational study. Setting: A univer sity hospital intensive care unit. Patients: Twenty-four critically il l patients with different risk factors for ARDS, Interventions: Arteri al and mixed Venous blood, as well as urine samples, were collected. I nvasive hemodynamic measurements were performed, Measurements and Main Results: Fifty nine Variables pertaining to the cardiorespiratory, he patic, immunologic, and renal systems and including plasma complement activation products C3a and SC5b-9 and polymorphonuclear elastase, wer e determined every 6 hrs for 3 days in patients at risk for ARDS. Asso ciations among variables were investigated and the predictive value of nonassociated variables for ARDS was determined. Patients who develop ed ARDS (n = 8) had lower white blood cell counts at the time they ent ered the study (p =.006) and during the first 24 hrs thereafter (p =.0 32). Also, plasma C3a concentrations were markedly higher during the f irst 24 hrs in patients who developed ARDS (p= .006), Plasma C3a had b etter predictive value than did white blood cell counts for cutoff poi nts set by discriminant analysis at 1075 ng/mL (1.075 x 10(-3) g/L) an d 5700 cells/mL, respectively. The combination of both variables in a discriminant function improved the predictive value for ARDS. Conclusi ons: The most notable and nonassociated alterations observed in patien ts who developed ARDS were lower white blood cell counts and higher pl asma C3a concentrations compared with counts and concentrations in pat ients who did not develop ARDS, Plasma C3a concentrations showed bette r predictive value than white blood cell counts. The combination of wh ite blood cell counts with plasma C3a concentrations synergistically i mproved the predictive value for ARDS, This combination may prove usef ul for identifying subpopulations at highest risk for ARDS and may con tribute to make treatment at an early stage of the syndrome possible.