Literature review > Issue_3 > Review Diemert et al. 

Enhanced diagnostic approaches for Neisseria gonorrhoeae, applicable either to an individual or to a population, can materially assist the management of gonorrhoea at a case or programmatic level. Nucleic acid amplification assays (NAAT) have a number of recognised advantages, including those of increased sensitivity of detection and ease of sample collection and transport, when compared with culture based methods. While many of the anticipated benefits of NAAT testing in the diagnosis of gonococcal disease have been realised in practice, increasing experience with NAAT over time and in different settings has tempered initial enthusiasm. It is now recognised that there is a need both for a considered approach to the application of NAAT under different conditions and for an awareness of their limitations. These considerations include, but are not limited to, those of sensitivity and specificity of primary and, where used, 'confirmatory' tests, the prevalence of infection in various populations and test parameter 'tuning' to deal with different 'signal/noise' phenomena as well as other issues of laboratory testing.

The study of Diemert et al of the utility of the COBAS AMPLICOR NG NAAT was undertaken in a low prevalence setting in Canada. The authors sought to extend observations on testing algorithms applied to the diagnosis of gonorrhoea by molecular methods, and in particular to further clarify the role of 'confirmatory' testing, in this instance by a 16S rRNA PCR. Although the commercial version of this latter test has now been withdrawn, the study provides some important insights into approaches required for testing for gonorrhoea by NAAT. The description of the study design is in general quite detailed, especially in regard to testing protocols. In describing the sample source however, the authors refer to a 'diverse array of outpatient clinics' to derive the samples. It is unlikely that the prevalence of disease would be uniform in all these groups and some of them would reasonably be expected to fall into a higher prevalence group. It would have been helpful if the disease prevalence in the sample subsets was available and if the source of positive samples by patient group was alluded to in the report.

The outcome of the study was presented mainly by means of an analysis of the sensitivity, specificity and PPV, which was the stated aim. Not surprisingly, the PPV was low, but especially so for samples from all sites from women. Additionally the CIs for the PPVs were wide ranging, probably because it was possible to 'confirm' positive AMPLICOR PCR tests by 16S rRNA testing in only 72 of 230 instances. Further, analysis through the use of culture as the 'gold standard' in the study was compromised by the availability of only 737 culture/PCR sample pairs, of which 7 were culture positive and 29, including these 7, PCR positive in the initial assay. Also of relevance was the increase in PPV observed as the OD readings were stratified within and above a defined equivocal zone. Although false positive reactions occurred more often at lower ODs this was not a universal finding. As a result, both here and in other cited studies of the reliability of this test in higher prevalence populations, it was not possible to arrive at an OD cut-off reading that would provide a satisfactory balance between sensitivity and PPV. In this study, the "high specificity was obtained using an algorithm that required retesting with OD readings that fell within a broad equivocal zone…"

There is little argument with the proposition in this and other studies that algorithms for retesting of samples positive by AMPLICOR PCR need to be in place and that these should be adapted to meet population requirements. A number of suggestions for dealing with the issue of retesting are published [1] and are pertinent, given the withdrawal of some commercial NAAT. These protocols come at additional cost, and, if patient recall is involved, logistic difficulty and patient anxiety. Further, while one can agree that the test evaluated here is highly sensitive, subsequent data provide less comfort in regard to estimates of specificity for this and also other NAAT [2]. This suggests that additional testing, even using a second assay directed at different targets, could not guarantee a positive result. Again, if the second assay is of lower sensitivity than the first NAAT, some true positives will fail to confirm on repeat testing. With these caveats in mind, it is perhaps best if the second assay is regarded as 'supplemental' rather than truly 'confirmatory'.

Additional, mainly technical and methodological, differences between NAAT also exist. Moves towards international standards for mandatory internal amplification controls have begun [3]. In the area of STI NAAT, the issue of specimen integrity has also been raised. Further, some assays, notoriously those based on the use of the cppB gene as the target, have a variable sensitivity according to the prevalent gonococcal population. It should also be remembered that the above remarks are based on studies confined to examination of urine or genital tract samples. Finally, it is acknowledged that newer 'real time' assays may be less sensitive than the NAAT which was the basis of this report [4].

Recognition of these issues does not devalue the utility of NAAT in STI. Rather, their resolution, to the extent that this is possible, allows for a proper appreciation of the value as well as the limitations of NAAT so that a valid and reliable laboratory result can be produced and properly interpreted within a relevant clinical context.

References
1. Centers for Disease Control and Prevention. Screening tests to detect Chlamydia trachomatis and Neisseria gonorrhoeae - 2002. MMWR 2002;51(No.RR-15)

2. Palmer HM, Mallinson H, Wood RL, Herring AJ. Evaluation of the specificities of five DNA amplification methods for the detection of Neisseria gonorrhoeae. Journal of Clinical Microbiology 2003;41:835-837.

3. Hoorfar J, Cook N, Malorny B, Wagner M, de Medici D, Abdulmawjood A, Fach P. Making internal amplification control mandatory for diagnostic PCR. Journal of Clinical Microbiology 2003;41:5835.

4. Whiley DM, LeCornec GM, Mackay IM, Siebert DJ, Sloots TP. A real-time PCR assay for the detection of Neisseria gonorrhoeae by LightCycler. Diagnostic Microbiology and Infectious Diseases 2002;42:85-89.

about SDI | newsletters | grants | publications | literature reviews

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