Antinuclear antibody testing: discordance between commercial laboratories

Antinuclear antibody (ANA) test results frequently affect the course of patients' evaluations, diagnosis, and treatment, but different laboratory centers may yield conflicting results. This study investigated the degree of agreement between laboratory results in a group of subjects who had ANA testing performed at two commercial laboratories. This was a chart review study, in which all ANA tests ordered by the authors from one commercial laboratory over a 4-year period were queried. Corresponding patient charts were reviewed, and if ANA testing had also been performed at the second commercial laboratory, subjects were entered into the study. The primary measurement was agreement between paired ANA results, and we performed sensitivity analysis using varying criteria defining agreement (criteria A to criteria D [strictest to most lenient definition of agreement]). Other data captured included relevant data obtained through the course of evaluation (e.g., presenting complaints, exam findings, other laboratory data) and final diagnoses. Of 101 paired ANA tests, there was 18 % agreement according to the strictest criteria and 42 % according to the most lenient. Of the seven subjects with ANA-associated rheumatic disease, none of the paired tests were in agreement according to criteria A (two agreed according to criteria D). Our findings demonstrate poor agreement between paired ANA tests performed at two commercial laboratories. The low level of agreement may have far-reaching clinical implications. Specifically, this finding calls into question the reliability of ANA testing as it is currently performed and suggests that results may in part depend upon the laboratory center to which patients are referred.