An analysis of discrepancy between point-of-care and central laboratory international normalized ratio testing in ED patients with cerebrovascular disease.

Objective: Our objective was to identify demographic, clinical, and operational variables associated with discrepancy between point-of-care (POC) and central laboratory international normalized ratio (INR) results in emergency department (ED) patients with acute cerebrovascular disease.

Methods: We conducted a retrospective, observational cohort study of a series of 637 patients with acute cerebrovascular disease over 30 months who underwent simultaneous POC, using the i-STAT POC analyzer (Abbott, Princeton, NJ), and central laboratory INR testing at ED presentation. Point-of-care INR results greater than ± 0.25 INR units from the central laboratory INR value were considered discrepant. We analyzed potential predictors of POC INR discrepancy from demographic, clinical, and operational variables using multivariable logistic regression. We evaluated the change in POC INR discrepancy incidence over the study interval using analysis of variance methodology.

Results: The final diagnoses of the 637 subjects were acute ischemic stroke (n=427), transient ischemic attack (n=105), and intracranial hemorrhage (n=105). Discrepant POC INR results occurred in 21.5% (137/637) of subjects. The mean bias between POC and central laboratory INR was 0.24 ± 0.69 (range, 0-11.3). Significant covariates of POC INR discrepancy were oral anticoagulant use (odds ratio, 3.03; confidence interval, 1.37-6.68) and increasing activated partial thromboplastin time (aPTT) (odds ratio, 1.07; confidence interval, 1.02-1.12). We observed a significant reduction trend in the incidence of POC-central laboratory discrepancy over the study period, decreasing on average at 0.42% per month (F=5.59, P=.025).

Conclusion: In this retrospective study, oral anticoagulant use and increasing aPTT were significantly associated with POC INR discrepancy in ED patients with acute cerebrovascular disease. Point-of-care INR discrepancy incidence decreased over the study interval.