|
|
||||||||||
 |
 |
Literature review > Issue 1 > Review Page-Shafer et al. |
||||||||
|
|
The authors undertake a relevant research question about the convenience of trying a nucleic acid amplification test (LCR) for the diagnosis of Neisseria gonorrhoeae from the throat of infected men, as opposed to the classical choice of bacterial culture in selective medium. The DNA amplification method detected 9.5% true prevalence, after GC pilin gene LCR discrepancy analysis, of pharyngeal N. gonorrhoeae infection (not 11%) in comparison to 4.5% by culture. The authors state they utilized the confirmatory pilin LCR only for pharyngeal swabs that were culture negative/ LCR positive, but apparently they also performed the assay for the only sample that was culture positive/LCR negative (according to table 2). The question arises then why they did not test with the pilin LCR the remainder of the culture positive samples. The authors show that the approach of molecular diagnosis of gonococcal infection among homosexual men suffering pharyngeal gonorrhea due to fellatio is feasible and more effective than bacterial culture in terms of a higher sensitivity, although with a somewhat reduced specificity. This was an expected result and even though LCR has not been standardized nor recommended by the manufacturer for use with pharyngeal swabs, this report is a starting point for replicating the procedure elsewhere and confirming its usefulness, especially in the clinical setting that would benefit from identifying and treating more cases among high risk MSM. Actually, an important point to mention at this stage is that LCR is going off the market shortly because of the manufacturer's decision. Therefore, if further validating studies like this are to be performed they will require a different nucleic acid amplification test. The authors acknowledge the limitations of their study in regard to sampling design, sample size, and realizing that LCR only positive results may belong to subjects who are no longer infectious. However, they missed the point by not recognizing that LCR (or any other nucleic acid amplification test) is a resource not readily available in most developing settings where MSM may request STI diagnosis. Thus, the need remains to develop and make available rapid and reliable STI diagnostic approaches that require no major training nor the use of laboratory equipment, have a lower cost, can withstand relatively high humidity and temperature conditions, and are user-friendly [1-3]. In any event, it should be kept in mind that pharyngeal gonorrhea diagnosis by nucleic acid or antigen detection does not address the antimicrobial susceptibility of gonococci strains. In this regard, it would have been interesting if the authors had mentioned the treatment given to the infected men and the outcome after follow up [4]. Granted, these were not the aims of the study. However, the consideration is posed because of the emphasis the authors put on the targeted screening of pharyngeal gonorrhea among high risk populations in order to help reduce the burden of gonococcal disease. References:
Most studies of pharyngeal gonorrhea infection have relied on culture to detect Neisseria gonorrhoeae. However, culture may have relatively low sensitivity for these extragenital infections. Thus, results on the performance of more sensitive nucleic acid amplification tests, such as ligase chain reaction (LCR), are important and a welcome addition to our current knowledge. The data presented by Page-Shafer et al. show that, prior to discrepancy resolution, LCR has a sensitivity of 89% relative to culture and a specificity of 92.6%. A second LCR test (referred to as LCR2 below) is used to resolve the status of 14 culture negative, LCR positive specimens. Following resolution, a "gold standard" is created in which all culture positives and all culture negatives that were positive by LCR2 are defined as positive. The authors report a sensitivity of 94.7 % and a specificity of 97.8% relative to this gold standard. This is a standard discrepancy analysis approach and is known to bias the estimate of sensitivity upwards [1]. One simple way to reduce the bias is to test a sample of the culture negative, LCR negative ("negative-negative") specimens by LCR2 as well [2]. Other approaches have also been proposed [3]. Testing a sample of the negative-negative specimens should be routinely implemented whenever discrepancy analysis is used. In this particular study, however, bias in the sensitivity estimate may be secondary to the relatively wide confidence interval attributable to the low number (19) of gonorrhea positive individuals in the study. Additional data from such individuals is clearly needed to more precisely identify the sensitivity of LCR, a point the authors readily acknowledge. The authors also report the sensitivity, specificity and predictive values of culture relative to the constructed gold standard. However, these results are more problematic. Since the gold standard is defined, in part, by the culture test, this clearly violates the principle that the performance characteristics of a test should be evaluated relative to an independent gold standard. In fact, since all specimens that are culture positive are defined as positive by the gold standard, the specificity and positive predicted value of culture must be 100% and the confidence intervals reported in for these values are not meaningful. References: |
|||||||||
|
about SDI | newsletters | grants | publications | literature reviews WHO
Home -
WHO
Search - TDR Home - SDI Home -
SDI Contact us
|
||||||||||
.gif){kind=small}