Literature review > Issue 8 > Review on Katz et al. 

The remarkable sensitivity and specificity of nucleic acid amplification (NAAT) tests provides a powerful tool and an enticing lure to physicians to screen for sexually transmitted infections (STIs). Indeed, physicians do not test enough -- rates of testing are far lower than they should be in many settings and physicians are constantly urged to test more. Laboratory-based screening is often the only way to detect gonococcal and chlamydial infections since these STIs are often asymptomatic and syndromic management is sub-optimal in its sensitivity and specificity. Moreover, clinical teaching supports over-testing and even over-treating in the situation wherein a disease can cause great harm and testing and treating creates little risk. Thus, why not test everyone?

The report by Katz et al. is a reminder that there is a downside to knee-jerk STI testing. The report details an investigation of a cluster of five presumed false positive N. gonorrhoeae tests. As it turned out, all of these tests emanated from a single laboratory within a defined period of months. Based on the laboratory's report of the number of tests conducted and manufacturer's information on test characteristics, the authors calculated that in the setting of 1.06% of all gonococcal tests being positive, the predictive value of a positive test was only 60%. Said another way, the likelihood that a positive test was a true positive was 60% and the likelihood that a positive test was a false positive was 40%. This is not a trivial matter, as a false positive necessitates both index patient and partner treatment, with potentially adverse ramifications for the relationship between the two.

Strengths of this study are to re-iterate the truism that in very low risk populations, a positive test, even using a sensitive and specific assay, may well be a false positive. As noted in the accompanying editorial [1], "the ultimate responsibility for the application of screening tests rests with the physician." That is, the interpretation of a screening test must consider the pre-test probability of disease. That pre-test probability should rest upon consideration of the prevalence of gonococcal disease in a given locality, and the presence of risk, which include factors such as young age and non-monogamy [2]. The characterization of these risk factors requires a good sexual history.

The study also cautions the clinician about the use of bundled gonorrhea and chlamydia tests. In the Katz study, the positive predictive value of a chlamydia test was 89.4%, which is not terrible, while the predictiveness of a gonorrhea test was far lower. Thus, in settings wherein the prevalence of chlamydia substantially exceeds that of gonorrhea, screening for gonorrhea should not accompany screening for chlamydia.

Weaknesses of the study are: 1) the cluster of false positives came from a single laboratory and it is possible that specimen handling or human error may have been involved in the generation of false positive results; 2) the presumption of a false positive (rather than a true positive) may have been misclassified, although the definition used was reasonable, based on: a negative retest using a different NAAT and culture, who did not receive appropriate antimicrobial therapy or who received therapy but not long enough prior to re-test to account for a negative re-test, and with a low prior probability of infection; 3) the figures used in the calculation of positive predictive value were derived from literature published using other populations or from self-report (by the laboratory), which may have some margin for error.

All-in-all, the take home point is well-founded: take a sexual history, consider the background prevalence, and if the likelihood of disease is very low and the reason for screening is not compelling, either don't screen or use a confirmatory test.

References:

1. Klausner JD. The NAAT is out of the bag. Clin Infect Dis 2004;38:820-1.

2. Kohl KS, Sternberg MR, Markowitz LE, et al. Screening of males for Chlamydia trachomatis and Neisseria gonorrhoeae infections at STD clinics in three US cities. Int J STD & AIDS 2004;15:822-8.

about SDI | newsletters | grants | publications | literature reviews

WHO Home - WHO Search - TDR Home - SDI Home - SDI Contact us
(c) WHO/OMS 2001