Assessment of using fingerstick blood sample with i-STAT point-of-care device for cardiac troponin i assay

Purposes The purpose of this study is to compare fingerstick point-of-care (POC) testing for cardiac troponin I to conventional venipuncture POC testing using the i-STAT device. Basic procedures This study was conducted with institutional review board approval in the emergency department (ED) of a 535-bed suburban level I trauma center from June to August 2011. Fingerstick blood samples were collected from consenting patients for whom standard-of-care venipuncture POC troponin (POCT) testing had been ordered as part of their workup. Cardiac troponin I (cTnI) assays were performed using the i-STAT 1 device (Abbott Point of Care, Princeton, NJ). The data were subjected to categorical comparison, linear regression, and Bland-Altman agreement analysis using SAS 9.2 software (SAS, Cary, NC). Main findings Eighty-nine cTnI levels were measured by both fingerstick and standard venipuncture ED POC testing. Four resulted in cartridge error; the remaining 85 were analyzed. Fingerstick testing, compared with standard ED POCT, has a positive predictive value of 1.00 (0.48, 1.00), negative predictive value of 0.96 (0.89, 0.99), sensitivity of 0.625 (0.24, 0.91), and specificity of 1.00 (0.95, 1.00). The relationship between methods appears linear, with linear regression equation ED POCT level = 0.0062 + 1.3752* fingerstick level (P < 0.0001). Bland-Altman agreement analysis yielded a mean difference between fingerstick and ED POCT of - 0.0095 with limits of agreement of - 0.0625 to 0.0435. Principal conclusions Fingerstick cTnI testing using the i-STAT device is not accurate enough to determine the exact troponin level without the application of a corrective term. Fingerstick testing is, however, accurate in qualifying troponin levels as negative, borderline, or positive and is, therefore, capable of providing clinical information that may guide diagnostic and therapeutic decision making.