AN INFLAMMATION METER TO REVEAL THE PRESENCE AND EXTENT OF INFLAMMATION IN OLDER PATIENTS
Rivka Rotstein, David W Zeltser, Itzhak Shapira, Shlomo A. Berliner, Daniel Avitzour, Tzvi Dwolatzky, Nadir Arber · Journal of the American Geriatrics Society · 2000
To the Editor: Increasing number of older patients reside in nursing homes where real-time sophisticated laboratory services are not readily available. The prompt diagnosis of infection/inflammation and the determination of the intensity of the acute phase response are essential for the initiation of appropriate therapy. We have taken advantage of the fact that leukocytes1 and erythrocytes2 become sticky during infection/inflammation. We present a simple slide test and image analysis to reveal the number of leukocytes, their adhesiveness/aggregation state, and the aggregability of erythrocytes in a single drop of peripheral venous, citrated blood.3–5 We examined 118 patients older than age 66 (39 with pneumonia, 26 with urinary tract infection, 22 with soft tissue infections, 18 with sepsis, and 13 with other bacterial infections) and 129 controls. The principles of the preparation of the peripheral blood slides have been described previously.1 The image analysis system (INFLAMET™) consisted of a Pentium computer running Windows 95 equipped with a Matrox Meter color frame grabber, a color CCD camera, and a microscope that was operated at x200 magnification, resulting in an image resolution of 0.4 microns per pixel. Nine images were taken from each slide. The fields of view were chosen systematically to sample different regions on the slide. Each image was processed separately, and the outputs were then averaged to form the final slide outputs. The nine fields of view cover a total area of 0.6 mm2. By using this system, we could determine the number of leukocytes per mm2, their degree of adhesiveness/aggregation (LAA), and the degree of erythrocyte adhesiveness/aggregation (EAA). The white blood cell count (WBCC) was performed by using the Coulter STKS autoanalyzer, the erythrocyte sedimentation rate (ESR) by the method of Westergren, the fibrinogen concentration by the method of Clauss, and quantitative C-reactive protein (CRP) was performed by using laser nephelometry and specific anti-human CRP antibodies. Statistical analysis was carried out using the SPSS statistical package. A significant (P < .001) difference was noted between patients and controls using the conventional laboratory variables (WBCC and differential, ESR, CRP and fibrinogen) as well as those obtained using the Inflammation Meter (leukocytes per mm2, leukocyte adhesiveness/aggregation test (LAAT), and the erythrocyte adhesiveness/aggregation (EAAT)). The specific Pearson correlations for the entire cohort were as follows: between the number of leukocytes per mm2 and the WBCC, r = 0.83, P < .001, n = 245; between EAAT and the ESR, r = 0.56, P < .001, n = 237; between the EAAT and fibrinogen, r = 0.51, P < .001, n = 219; and between the CRP and the LAAT, r = 0.50, P < .001, n = 198. A discriminant analysis was performed to classify the individuals into one of two groups, either acute bacterial infection or controls. Included in this analysis were 86 patients with acute bacterial infections and 85 controls for whom the entire series of results of conventional laboratory and Inflammation Meter variables were available. By looking at the results presented in Table 1, one can see that the diagnostic yield of the variables obtained by the Inflammation Meter are similar to those obtained by the conventional tests. Representative pictures showing both a patient with Gram-negative sepsis and a control are seen in Figure 1. . A representative picture showing (top) a patient with a Gram-negaative sepsis and (bottom) a control. We conclude that it is feasible to obtain a “Picture of Inflammation” by using a simple slide test and imaging analysis. This picture is obtained by looking at the number of white blood cells per mm2, their degree of adhesiveness/aggregation, and the erythrocyte aggregability. We could also show that the diagnostic yield in terms of discriminating between patients with acute bacterial infections and controls, as obtained by our Inflammation Meter, is similar to that which can be seen using conventional markers of the acute phase response. The fact that our system can be operated by any paramedical personnel after a short training program and that images can be transferred easily via telephone or the Internet to a physician who is remote from the scene might be attractive for clinicians who work in small centers that are not equipped with routine 24-hour laboratory services. Further studies are in progress to delineate the appropriate clinical applications for this rapid diagnostic innovative approach.