Objectives To compare performance of the ResistancePlus kit (SpeeDx, Australia) with in-house methods for the detection of Mycoplasma genitalium-specific DNA and mutations associated with resistance to macrolide antimicrobials, directly from clinical specimens.
Methods Assay specificity and sensitivity was analysed using DNA from 46 non-M. genitalium organisms and standard curve analysis, respectively. A panel of archived DNA extracted from 97 M. genitalium-positive clinical specimens, for which the macrolide susceptibility genotype had been previously determined, were tested on the assay and results compared.
Results Final analytical specificity was 100%. Sensitivity was detected to at least 140 genome copies/µL. The assay detected M. genitalium in 92/97 (94.9%, 95% CI 88.4% to 98.3%) previously positive specimens. The genetic macrolide susceptibility assigned was concordant with previous results in 85/92 (92.4%, 95% CI 85.0% to 96.9%) specimens or 85/97 (87.6%, 95% CI: 79.4% to 93.4%) when the false-negative specimens were included. On seven (7/92, 7.6%) occasions, resistant specimens were called susceptible. Further testing resolved discrepancies for all but five (5.2%) specimens.
Conclusions The ResistancePlus assay generally performed well in comparison to methods currently employed at the reference laboratory. It detected a range of different mutations; however, a small number of specimens that were genotyped as macrolide resistant by Sanger sequencing were either not detected by the assay or were genotyped as susceptible. This could impact on treatment outcomes if assay results were used for patient management.
- antibiotic resistance
- molecular techniques
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Mycoplasma genitalium is a cause of urethritis in men and cervicitis in women1 and has been increasingly associated with upper genital tract infection and pelvic inflammatory disease.1 2 In patients with non-gonococcal urethritis (NGU) and in the absence of M. genitalium diagnostic testing, UK guidelines3 recommended treatment with 1 g azithromycin until earlier this year. M. genitalium is slow growing, highly mutagenic4 and replicates both intracellularly and extracellularly.5 Azithromycin has a much longer intracellular than extracellular half-life,5 so a 1 g single dose can provide selection pressure resulting in azithromycin-resistant mutants with chromosomal point mutations at positions 2058 and 2059 (Escherichia coli numbering) of region V of the 23S rRNA gene. The European M. genitalium and updated UK NGU treatment guidelines6 7 recommend prescription of an extended 5-day regimen (500 mg azithromycin followed by 250 mg azithromycin daily for 4 days). This does not appear to select for resistance in M. genitalium to the same extent,1 7 likely due to the intracellular and extracellular tissue concentrations both remaining above the therapeutic threshold for longer. Where diagnostic tests are available, M. genitalium-positive specimens should ideally be tested for macrolide resistance-associated mutations to aid patient management.1 Hence, we evaluated the M. genitalium ResistancePlus kit (SpeeDx, Australia), which simultaneously detects M. genitalium-specific DNA and mutations that have been associated with resistance to macrolide antibiotics.8
The ResistancePlus assay utilises a novel real-time PCR technology, PlexPCR, for the specific amplification of M. genitalium and concurrent detection of macrolide resistance-associated mutations in DNA extracted from clinical specimens.8 Master mix was prepared according to the manufacturer’s instructions, and reactions contained 5 µL template DNA. Assays were run on an LC480 II (Roche, UK) using manufacturer-recommended reaction conditions. Results were interpreted using the ResistancePlus MG Analysis Software LC480 V.1.0.
Sensitivity and specificity
Assay sensitivity was defined using serial dilutions (1.4 ng/µL–0.000014 ng/µL), tested in triplicate, of an M. genitalium genomic DNA (gDNA) extract of ATCC strain 49895. To test specificity, an exclusivity panel of 46 non-M. genitalium strains was tested: Mycoplasma hominis (n=1), Ureaplasma urealyticum (n=1), U. parvum (n=1), Chlamydia trachomatis (E-L3) (n=12), Neisseria gonorrhoeae (n=7), Haemophilus ducreyi (n=1), Herpes simplex virus 1 (n=1), Herpes simplex virus 2 (n=1), Treponema pallidum (n=1), Trichomonas vaginalis (n=1), Moraxella lacunata (n=1), M. catarrhalis (n=1), N. cinerea (n=1), N. elongata (n=1), N. lactamica (n=1), N. meningitidis (n=1), Candida albicans (n=1), E. coli (n=1), Gardnerella vaginalis (n=1), Lactobacillus jensenii (n=1), L. crispatus (n=1), L. gasseri (n=1), L. iners (n=1), L. vaginalis (n=1), Staphylococcus aureus (n=1), S. epidermidis (n=1), Enterococcus faecium (n=1), E. faecalis (n=1), Shigella flexneri (n=1).
M. genitalium-positive panel
A panel of archived DNA sample extracts from 97 M. genitalium-positive clinical specimens (collected and originally tested between 2010 and 2013, and reconfirmed as M. genitalium-positive prior to initiation of this study, using a modified in-house real-time PCR (RT-PCR) targeting the M. genitalium MgPa adhesion gene9) was tested on the ResistancePlus assay. The 23S rRNA sequence at positions 2058 and 2059 (E. coli numbering) for each sample had been determined previously by Sanger sequencing. Discrepant samples were retested with the ResistancePlus assay and any remaining discordant specimens were resequenced.
Sensitivity and specificity
The lowest, repeatedly detectable dilution of M. genitalium ATCC 49895 gDNA by the ResistancePlus assay corresponded to 0.00014 ng/µL (140 fg/µL) or approximately 140 genome copies/µL. Forty-five strains from the exclusivity panel were classified as negative by the assay interpretive software. A false-positive signal was recorded by the LightCycler software for one C. trachomatis sample, serovar L1, where the fluorescent plot was atypical. The result was over-ruled by Public Health England users, resulting in 100% assay specificity. Without over-ruling, autocalling by the software resulted in a specificity of 97.8%.
Detection of M. genitalium
The ResistancePlus assay detected M. genitalium in 92/97 specimens that were previously positive; 5/97 (5.2%) specimens were incorrectly identified as negative for M. genitalium. The sensitivity of the M. genitalium detection aspect of the ResistancePlus assay was 94.9% (92/97).
Detection of macrolide susceptibility genotype
The macrolide susceptibility genotype was concordant with the original sequencing result in 85/92 specimens recognised as positive for M. genitalium by the ResistancePlus assay (table 1) resulting in 92.4% genotyping sensitivity, with seven false-negative results (ie, ‘SpeeDx-Susceptible(S)/in-house-Resistant(R)' specimens). If the five detection false-negative specimens were included in this calculation, sensitivity was 87.6% (85/97). Resistance mediated by a range of different clinically relevant mutations was detected by the ResistancePlus assay although it is not designed to report on the specific mechanism found (table 1).
Discrepancy testing and analysis
Where discordant results were observed for M. genitalium and/or resistance detection, the samples were retested on the ResistancePlus assay (table 1). Where the discrepancy was for detection of M. genitalium in the specimen (positive by in-house MgPa RT-PCR, negative on the ResistancePlus assay), repeat testing resolved the issue in 4/5 samples with only one sample repeatedly giving a false-negative result on the ResistancePlus assay. In the four samples that tested positive on retest, the macrolide susceptibility genotype assigned by the ResistancePlus assay was concordant with the original sequencing results.
Seven specimens that tested positive for M. genitalium on the ResistancePlus assay were assigned macrolide susceptibility genotypes that were discordant with the original sequencing results. In all cases, sequencing of the 23S rRNA gene had indicated a resistant genotype; however, the ResistancePlus assay assigned a susceptible genotype. This included one sample where the in-house sequencing results had indicated a mixed sample with both A2059G and wild-type sequences. With repeat testing on the ResistancePlus assay, 3/7 samples were assigned resistant genotypes and so became concordant with the sequencing results (including the mixed sample), while 4/7 remained discordant (2/7 were reassigned susceptible genotypes and 2/7 were negative on retest) (table 1).
In order to resolve the discrepancy seen in these four specimens (two ‘SpeeDx-S/in-house-R’ specimens and two ‘SpeeDx-negative/in-house-positive’ specimens), the 23S rRNA gene was resequenced. In two of the specimens, the repeat sequencing confirmed the initial sequencing result (resistant). In one specimen, a mixed population was found where both susceptible and resistant organisms were present (indicated by the presence of both A and G nucleotides at position 2058). The final specimen failed to amplify. The sensitivity of the genotyping assay was recalculated as 95.8% (92/96) (not including the remaining false-negative, 94.9% including it (92/97)).
Overall, the M. genitalium ResistancePlus Kit (SpeeDx, Australia) performed well when compared with an in-house RT-PCR for detection of M. genitalium. The commercial assay was able to detect M. genitalium in 94.9% (92/97) of the in-house PCR-positive specimens, and this increased to 98.9% (96/97) with repeat testing. In addition, the assay was 100% (46/46) specific for M. genitalium when tested against an exclusivity panel. As specimens were tested retrospectively, requiring storage, DNA degradation cannot be excluded, and could have adversely impacted on the assay performance.
The commercial assay gave concordant macrolide susceptibility genotypes with previous PCR amplicon-sequencing results for 92.4% (85/92) of specimens. With repeat testing, this increased to 94.9% (92/97) (one false-negative detection result and four discrepant genotype results remained). The assay detected clinically relevant mutations at both the 2058 and 2059 positions of the 23S rRNA gene, although it did not distinguish between them, reporting specimens only as susceptible or resistant. There did not appear to be a bias as to where discordant results were seen.
Two specimens appeared to contain mixtures of macrolide-susceptible and macrolide-resistant M. genitalium strains, since either A2058G or A2059G mutations were present together with wild-type sequences. On initial testing with the ResistancePlus assay, both samples were assigned susceptible genotypes. Repeat testing resulted in one of the samples being reassigned a resistant genotype, but the other sample remained susceptible. These samples would be particularly challenging to identify by any means, demonstrated by the fact that one was only identified through resequencing to resolve the discrepant results.
There were no instances where samples that were genotyped as susceptible by sequencing were called resistant by the ResistancePlus assay, meaning no unnecessary prescription of the second-line antimicrobial, moxifloxacin. Discordant results between the in-house results and the commercial assay were observed for seven specimens where sequencing had denoted the strain to be carrying a mutation and therefore to be macrolide resistant. If the ResistancePlus assay had been used to guide patient management, these errors (‘SpeeDx-S/in-house-R’) would indicate, incorrectly, that the first-line therapy (azithromycin) could be prescribed when in fact this treatment would be unlikely to clear these infections.
Overall, the results reported in this evaluation were in agreement with previously reported sensitivity and specificity rates for this assay10; however, we found the assay to be more sensitive for macrolide-genotype detection than Le Roy et al (2017).11
To conclude, the M. genitalium ResistancePlus Kit generally performed well against in-house assays to detect M. genitalium and infer macrolide resistance. However, a number of strains were not detected or were incorrectly genotyped on initial testing. While in the reference laboratory setting it was possible to identify discrepant results in this previously characterised strain panel, this would not be possible in a routine laboratory working on prospective specimens which may impact on the assay’s performance. In light of increasing reports worldwide of high levels of macrolide resistance in this organism, use of a test such as the one evaluated here would provide important information to the clinician in real-time allowing more appropriate prescribing.
Handling editor Nigel Field
Contributors RP carried out the laboratory work and prepared the first draft of the manuscript for this evaluation. MJC, HF and NW contributed to the analysis of results and subsequent drafts of the manuscript.
Funding This evaluation was funded and commissioned by SpeeDx. Funding was provided by SpeeDx for RP to attend and present this work at ECCMID 2017.
Provenance and peer review Not commissioned; externally peer reviewed.