Performance characteristics of platelet autoantibody testing for the diagnosis of immune thrombocytopenia using strict clinical criteria

Misclassification of immune thrombocytopenia (ITP) is common, which might undermine the value of platelet autoantibody testing. We determined the sensitivity and specificity of platelet autoantibody testing using the direct antigen capture assay for anti-glycoprotein (GP) IIb/IIIa or anti-GPIbIX in patients with 'definite ITP', defined as those with a documented treatment response. Sensitivity of platelet autoantiboody testing increased from 48 center dot 3% [95% confidence interval (CI) 43 center dot 5-53 center dot 2] for all ITP patients to 64 center dot 7% (95% CI 54 center dot 6-73 center dot 9) for definite ITP patients. Specificity was unchanged [75 center dot 3% (95% CI 67 center dot 5-82 center dot 1)]. High optical density values (>0 center dot 8) improved the specificity of platelet autoantibody testing but lowered sensitivity. In patients with a high pretest probability, platelet autoantibodies can aid in the diagnosis of ITP and may be most prevalent in certain patient subsets.

Automated summary

Misclassification of immune thrombocytopenia (ITP) is common and can undermine the value of platelet autoantibody testing. Sensitivity of platelet autoantibody testing is increased in patients with 'definite ITP', while specificity remains unchanged. High optical density values improve the specificity of platelet autoantibody testing but decrease sensitivity, suggesting that these antibodies may be most prevalent in certain patient subsets.