Related Experiment Videos
Short-term intraindividual variability in hemostasis factors. The ARIC Study. Atherosclerosis Risk in Communities
L E Chambless1, R McMahon, K Wu
1Department of Biostatistics, University of North Carolina, Chapel Hill 27514.
Abstract:
Recent epidemiologic studies found that there is a strong association of hemostatic factors with ischemic heart disease. The Atherosclerosis Risk in Communities (ARIC) Intraindividual Variability (IIV) Study was conducted to estimate the various components of variation in hemostasis factors measured in the ARIC Study and to estimate the measures of repeatability of these factors. A total of 39 subjects (16 men, 23 women) were studied. Each had blood collected three times, with a 1- to 2-week interval between each visit. The contributions of between-person variability, within-person (biologic) variability, and processing and assay variability were estimated. Then the reliability coefficient R was estimated as the proportion of total variance accounted for by between-person variance. The reliability coefficient can be interpreted as the correlation between measures made at repeat visits. Among the various analytes, the reliability coefficients were quite high for activated partial thromboplastin time and plasma factor VIII (R = 0.92, 0.86, respectively). Low repeatability was obtained for antithrombin III activity and protein C (R = 0.42, 0.56, respectively). The lack of repeatability for these variables derives mostly from the processing (field center and laboratory) variation. Other analytes--fibrinogen, plasma factor VII, and von Willebrand factor--were intermediate in repeatability. In comparing the analyte-specific high-level to low-level groups, no substantial difference of within-person plus method coefficient of variation between the two groups was found for any analyte except for factor VIII, whereas the corresponding variance components for most analytes were higher for the higher analyte level. Reliability coefficients from this ARIC IIV study are generally higher than those found in other studies, and this is related to the relative variations in populations studied and to the time between measurements.