Concordance between nucleic acid amplification technique and culture for the diagnosis of gonorrhoea

BACKGROUND

Nucleic acid amplification techniques (NAAT) for the detection of Neisseria gonorrhoeae facilitate non-invasive testing for gonorrhoea and may identify cases of gonorrhoea missed by culture alone. ¹ Owing to their acceptability and ease of transport, NAAT are increasingly being used in genitourinary (GU) and non-GU settings, although use in low-prevalence settings decreases the positive predictive value of the test. ² NAAT should be confirmed with microbiological culture as NAAT testing is unable to provide information on antimicrobial sensitivity. ³

NAAT testing and culture in GU clinics reportedly have a concordance rate of >90%. ^1,4–12 We present an evaluation of the concordance between NAAT and culture in a GU medicine clinic in East London.

METHOD

Between 1 April 2007 and 31 January 2008, clients attending the Department of Sexual Health, Homerton Hospital were tested for N. gonorrhoeae infection using the Becton Dickinson ProbeTec strand displacement assay (Becton Dickinson, Oxford Science Park, Oxford, UK). Male clients had a first void urine sample sent for NAAT testing. Female clients had a self-taken vulval swab sample if asymptomatic, or an endocervical sample if symptomatic. Positive NAAT tests had an endocervical or urethral culture plated on Philips medium prior to treatment.

Eligible specimens were any samples testing positive for N. gonorrhoeae on NAAT and/or culture. Samples from extragenital sites were excluded, as were samples with no contemporaneous NAAT and culture.

Discrepant samples were divided into those at extremely high risk of N. gonorrhoeae (e. g. gonorrhoea identified by microscopy, symptoms and signs suggestive of gonorrhoea and/or sexual contact of gonorrhoea) and those at moderate risk of gonorrhoea.

As a result of interim analysis, two further interventions were introduced:

From 1 July 2007 samples with discrepant results between NAAT and culture had both tests repeated prior to treatment;

RESULTS

A total of 164 male and 158 female samples tested positive for N. gonorrhoeae on NAAT and/or culture. Of these, two were excluded as they were from extragenital sites and 26 because contemporaneous culture and NAAT were not sent prior to treatment. The remaining results are shown in Table 1.

Among male patients, 23/159 (13%) were NAAT positive and culture negative. Of these 9/23 (40%) were clinically at a very high risk of having gonorrhoea, e.g. they had gonorrhoea identified by microscopy, were sexual contacts of gonorrhoea or had signs highly suggestive of gonorrhoea. Two out of 159 were NAAT negative and culture positive.

Among female patients 36/135 (27%) were NAAT positive and culture negative, of whom seven (20%) were at extremely high risk of having gonorrhoea. Four out of 36 (11%) discrepant results were self-taken vulval swabs, compared to 9/99 (9%) of the concordant samples (P > 0.7, Fisher's exact test). One out of 135 was positive on culture but negative on NAAT. This was a self-taken vulval swab.

After interim analysis, the decision was made to retest all discrepant samples prior to treatment. Nineteen of the 62 discrepant samples were retested. In 12 of these (63%), the NAAT had spontaneously reverted to negative.

CONCLUSIONS

There was concordance between the NAAT and culture results in 84% of male and 67% of female samples. Of the 62 discrepant results, 16 were at extremely high risk of gonorrhoea. Three samples were positive on culture but negative on NAAT. Two-thirds of discrepant NAAT results were negative when NAAT was repeated prior to treatment. A change in laboratory did not significantly affect these findings.

This study differs from the findings of previous studies, which found a concordance rate of 80–92%. ^1,12 These studies are taken in areas of much lower prevalence of gonorrhoea, which may contribute to the greater discordance rate.

The fact that three samples were negative on NAAT testing but positive on culture may imply inadequate NAAT sampling technique. There was no significant difference in the discordance rate between self-taken swabs and physician-taken swabs.

The initial analysis suggested that many of the discrepant results were false-negative culture rather than false-positive NAAT, as a quarter were at extremely high risk of having gonorrhoea. However, the fact that 10% of discrepant results were repeated and found to have spontaneously reverted to negative implies a higher rate of false-positive NAAT. Other studies have noticed a similar effect on repetition of discrepant results. ¹²

Other studies have highlighted that differences in collection and transport of specimens may affect sensitivities of cultures and thus affect comparison of culture results with corresponding NAAT. ⁸ In low-prevalence areas, the positive predictive value of nucleic acid tests may be <80%. ^2,10 This highlights the importance of consideration of results within the clinical picture.

As nucleic acid testing is used increasingly in settings with a lower prevalence of gonorrhoea, the positive predictive value of NAAT testing may fall further.