In Reply: Drs Plebani and Basso rightly point out that IgA-tTG tests using human antigens have been shown to have better diagnostic performance compared with older tests using guinea pig tissue. We presented information on test characteristics but did not carry out subgroup analyses to demonstrate their effects on diagnostic performance. The number of studies using guinea pig antigens is too small to compute pooled estimates, but bivariate analyses limited to studies using human tissue for IgA-tTG tests (n = 4) show a pooled estimate of 0.94 (95% confidence interval, 0.74-0.99) for sensitivity and 0.98 (95% confidence interval, 0.96-0.99) for specificity, a slightly better sensitivity and similar specificity compared with the overall pooled estimates presented in our review.
Danielle van der Windt, Frank Kneepkens, Henriette van der Horst. Diagnostic Testing for Celiac Disease—Reply. JAMA. 2010;304(6):639–640. doi:10.1001/jama.2010.1110