Interlaboratory test discordance in screening adolescents for Chlamydia with direct immunofluorescence

Chlamydia testing using direct fluorescent antibodies (DFA) is widely employed for screening and in clinical practice. We submitted duplicate DFA specimens to two pairs of laboratories during two study periods separated by a year. We found discordance between reports (one positive, one negative on the same individual) in 75% of chlamydia-positive cases when one pair of laboratories was compared and in 50% of cases when a second pair was compared. Calculations show that with optimum published test performance, a discordance rate of 40% might be anticipated. Our observed discordance of 75% suggests inaccurate test performance by one laboratory with regard to sensitivity and specificity, albeit not to levels differing from those reported by others for in-the-field test performance. The implications of suboptimal test sensitivity and specificity for epidemiologic study and clinical practice are discussed. Because a diagnosis of sexually transmitted disease may be disruptive for adolescent relationships, maximum laboratory accuracy is needed.