Article Text

Download PDFPDF

High prevalence of macrolide resistant Treponema pallidum strains in a London centre
  1. Craig Tipple,
  2. Myra O McClure,
  3. Graham P Taylor
  1. Jefferiss Trust Laboratories, Wright-Fleming Institute, Imperial College London, London, UK
  1. Correspondence to Dr Craig Tipple, Clinical Research Fellow, Jefferiss Trust Laboratories, Wright-Fleming Institute, Imperial College London, Norfolk Place, London W2 1PG, UK; c.tipple{at}


Objectives Macrolide resistant Treponema pallidum strains, caused by mutations in the 23S ribosomal RNA (23S rRNA) gene, are widespread and increasingly prevalent. The authors aimed to establish the strain types of T pallidum isolated from patients in a London sexual health centre and to determine the frequency of macrolide resistance.

Methods T pallidum DNA from blood and ulcer samples were subjected to strain typing and mutation analysis using previously described methods.

Results 18 samples were tested and a 23S rRNA point mutation conferring macrolide resistance was found in 66.6%. All resistant strains were collected from men who have sex with men and both the A2058G and the A2059G mutations were found. Two strain types were identified (14d/g and 14d/f); the predominant strain type was 14d/g and an association was noted between tp0548 type g and macrolide resistance.

Conclusions High levels of T pallidum macrolide resistance are present in London, UK, and this has clear implications for national treatment guidelines.

  • Treponema pallidum
  • syphilis
  • antibiotic resistance
  • molecular typing
  • back disorders
  • ergonomics
  • epidemiology
  • retrospective exposure assessment
  • male reproduction

Statistics from

Request Permissions

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.


British guidelines for the management of syphilis include macrolide antibiotics as alternative treatments for incubating and early disease.1 This advice comes with the caveat that intrinsic macrolide resistance is present in some strains of Treponema pallidum, although the UK prevalence is unknown.

Following trials showing comparable efficacy to benzathine penicillin, single-dose azithromycin was adopted for the treatment of incubating and early disease.2 3 This was followed by reports of treatment failures, mostly among men who have sex with men (MSM), attributed to a single point mutation in the T pallidum 23S ribosomal RNA (23S rRNA) gene, a mutation that is now geographically widespread and increasingly prevalent.3 4

In addition to the detection of resistance mutations, molecular analysis of T pallidum allows the identification of different strains and while there are common types, considerable diversity within epidemics has been demonstrated.5

In this study, we present genotypic evidence of macrolide resistance and strain typing data from London, UK.


Study participants and specimens

Samples were collected from patients attending St Mary's Hospital, London (July 2006–January 2008) in a cross-sectional study of T pallidum infection following ethical approval from the Hounslow and Hillingdon Research Ethics Committee.6

23S rRNA mutation detection

As previously described, a section from each copy of the T pallidum 23S rRNA gene (Genbank ref. AF200367) was PCR-amplified from extracted DNA.4 Separate aliquots of each amplicon were digested using the restriction endonucleases Mbo-II and Bsa-I and the resultant digestion pattern was analysed.4 7 For confirmation, the gene sequence of a third (undigested) aliquot was determined using the Applied Biosystems dRhodamine kit and 3100 capillary sequencer.

Plasmids containing DNA from Street 14 strain (an A2058G mutant obtained from a patient who failed macrolide therapy) or UW228B (A2059G mutant) constituted positive controls for both the initial PCR and the enzymatic digestion.

Strain typing

Identification of strains was based on the arp, tpr and tp0548 genes, as described.4 7 To summarise: following PCR amplification, the size of the arp amplicons, determined by agarose gel electrophoresis, is compared with Nichols strain T pallidum, which is known to contain 14 repeats of a 60 base pair sequence. The number of repeats forms the first part of the type (14, 15, etc). The second part requires Mse-1 restriction digestion of the tpr amplicon for comparison of the resulting digestion pattern with standard profiles. Each pattern is assigned a letter (a, b, etc). Finally, the sequence of the tp0548 gene amplicon is compared with one of a number of previously established sequences (a, b, etc).


T pallidum from 18 patients was available for analysis. All subjects were male, median age 37 years (range 24–54) and 61.1% were HIV-1 positive (table 1). Where sexual orientation was stated, 94.1% were MSM.

Table 1

Demographic, typing and mutation data

23S rRNA mutation analysis

23S rRNA mutations were identified in 12/18 (66.6%) samples of which 11 had the A2058G substitution (table 1). All 23s rRNA mutations were present in both copies of the gene.

Strain subtype analysis

Twelve samples had sufficient DNA remaining for strain-type analysis, resulting in the identification of five full and seven partial strain types (table 1). Where analysis of arp and tpr genes succeeded, only the 14d type was detected. Sequence analysis of the tp0548 gene classified these as 14d/g (four isolates) and 14d/f (one isolate) strains.

In total, 7 of the 12 tp0548 gene sequences were ‘g’ and the remaining 5 were ‘f’ of which 4 had wild-type 23S rRNA sequences and 1 was an A2058G mutant.

No association was identified between HIV-1 status and either full T pallidum subtype or type according to tp0548 sequence alone.


White MSM account for around three-quarters of syphilis cases in the UK and HIV-1 co-infection (34%) is common.8 Patients with early syphilis participating in this study (2006–2008) were also predominantly white MSM (94.1%); 61.1% were HIV-1 positive and two-thirds of the T pallidum isolates had genotypic evidence of resistance to macrolide antibiotics. These findings, the first such data from the UK, have implications for syphilis treatment guidelines.1

In San Francisco, the prevalence of resistant strains increased from 4% in 2000–2002 to 56% in 2004. This coincided with guidelines that included the use of azithromycin as an alternative for the treatment of incubating and subsequently early syphilis.3 While azithromycin has never been the preferred therapy for the management of syphilis in the UK, it has been increasingly prescribed for the treatment of non-specific urethritis and chlamydia. Samples prior to 2006 were not available to assess changes in the prevalence of resistant strains.

The suggestion that a single resistant strain may enter a population and become endemic is supported by evidence that the 23S rRNA mutation in the Street 14 strain is stable, when passaged in untreated rabbits.5 9 In this scenario, a mutant strain would have been imported into the London MSM population where it became the predominant circulating strain. If all the 23S rRNA mutants were a single clone, we would expect only one subtype. However, subtype analysis from this and other studies presents a more complex picture. Four of the full subtypes identified were 14d/g and the fifth was 14d/f. The higher level of success achieved by typing T pallidum from ulcer samples than from blood probably reflects the relative bacterial loads.6 While the number of full subtypes identified in this study was limited, other molecular epidemiology studies of syphilis in Scotland and Canada have identified several subtypes within these populations.5 10 It seems likely, therefore, that neither these epidemics nor the London epidemic represent the spread of a single clone, and a significant role for antibiotic selection pressure is suggested. This is supported by epidemiological evidence that macrolide exposure in the 3 months prior to syphilis diagnosis is associated with a 30% increase in resistance.9

In summary, we have shown that genotypic macrolide resistance in T pallidum is common in MSM in London and, therefore, we recommend caution in the use of macrolides for the treatment of syphilis, particularly in this MSM population. Although limited, our typing data suggest that the syphilis epidemic in London is not a single clone of resistant T pallidum, raising the possibility that resistance mutations are developing de novo.

Key messages

  • In two thirds of syphilis cases seen at our London centre, the causative Treponema pallidum has a mutation conferring high-level macrolide resistance.

  • The recent increase in Treponema pallidum infections is driven by more than one strain.

  • Macrolide antibiotics should be avoided for the treatment of incubating or early syphilis, especially for gay men in London.


The authors thank Professor Sheila Lukehart, University of Washington, Seattle, for donation of samples and advice.



  • Linked article 049494.

  • Funding This study was supported by the Mason Medical Foundation and the Imperial College Healthcare Biomedical Research Centre.

  • Competing interests None.

  • Ethics approval This study was conducted with the approval of Hounslow and Hillingdon REC.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles