Evaluation of automatic and semiautomatic coagulation assay instruments

The precision of five instruments (Coag‐A‐Mate, Auto‐Fi, Coagulyzer, Electra 600 and Fibrometer) in performing activated partial thromboplastin time (PTT) assays were compared using three different coagulation reagent systems (Ortho, Dade, and General Diagnostics) utilizing normal and abnormal control reagents. There were wide variations in mean PTT values and co‐efficient of variation (CV) among all instruments and for each instrument utilizing different reagents. The smallest CV was noted for the Coagulyzer using Dade reagents, the Coag‐A‐Mate using Dade and General Diagnostic reagents and the Auto‐Fi using General Diagnostic reagents. Auto‐Fi values were always longer than those determined with the Fibrometer whereas the optical detection instruments usually gave shorter values. The range of normal values for the prothrombin time (PT) and PTT assay was recorded simultaneously on blood samples of 52 healthy donors utilizing the Coagulyzer with Ortho reagents, the Coag‐A‐Mate with General Diagnostic reagents and the Auto‐Fi with Dade reagents. No differences were noted between male and female donors. There was a poor correlation among PT results recorded by all instruments. For the PTT assay, results obtained with the Coag‐A‐Mate and the Auto‐Fi had a correlation of 0.84. The normal mean plus 2 or 3 standard deviations determined by the Auto‐Fi and the Coag‐A‐Mate were used to classify blood samples from 263 patients as normal or abnormal when assayed by each instrument. For the PTT assay, coincidence was attained in 91.2% of the samples and discrepancies (a blood classified abnormal by one instrument and normal by the other) were evenly distributed for both instruments. For the PT assay, coincidence occurred in 85.1% of the cases and there was a statistically significant trend (P<o.05) for the Auto‐Fi in classifying abnormal samples categorized normal by the Coag‐A‐Mate.