Article Text

Download PDFPDF

Original article
Molecular epidemiology of gonorrhoea in Wales (UK)
  1. Michelle J Cole1,
  2. Daniel Rh Thomas2,
  3. Stephanie A Chisholm1,
  4. Adel N Abdullah3,
  5. Humphrey Birley4,5,
  6. Ian Hosein6,
  7. Elizabeth Kubiak5,
  8. Diana White3,
  9. Iona Martin1,
  10. Catherine A Ison1
  1. 1Sexually Transmitted Bacteria Reference Laboratory, Health Protection Agency, London, UK
  2. 2Communicable Disease Surveillance Centre, Public Health Wales, Cardiff, UK
  3. 3Genitourinary Medicine Department, Royal Glamorgan Hospital, Llantrisant, UK
  4. 4Department of Genitourinary Medicine, Cardiff Royal Infirmary, Cardiff, UK
  5. 5Department of Genitourinary Medicine, Royal Gwent Hospital, Newport, UK
  6. 6Department of Microbiology, Queen's Hospital, Romford, UK
  1. Correspondence to Michelle J Cole, Sexually Transmitted Bacteria Reference Laboratory, Health Protection Agency, 61 Colindale Avenue, London NW9 5EQ, UK; michelle.cole{at}


Objectives After a trend of increasing incidence of gonorrhoea in the 1990s, by 2004 the incidence was declining in England, but continuing to increase in Wales. This prompted an investigation of the epidemiology of gonorrhoea in Wales to inform future prevention and control measures.

Methods As an extension to Gonococcal Resistance to Antimicrobials Surveillance Programme, between May 2005 and September 2006, 540 consecutive gonococcal isolates were collected from three microbiology laboratories in South Wales. Isolates were typed using Neisseria gonorrhoeae Multi Antigen Sequence Typing tested for susceptibility to therapeutic agents and demographic and behavioural data were collected retrospectively from patient notes.

Results 163 sequence types (STs) were identified in 475 N gonorrhoeae isolates from 502 patient episodes. The most frequently observed STs (>20 isolates) were: 2, 752, 471, 249 and 8, all of which were susceptible to the antimicrobial agents tested. A significant association between ST and sexual orientation was identified, the most frequently observed STs occurring in young (median age <25 years) heterosexuals. STs 147, 4, 1634 and 64 predominated in men who have sex with men.

Conclusions We confirm the existence of common STs across the UK, as well as identify a number of types that were novel to Wales. Discrete sexual networks were identified, the most localised being in young heterosexuals. Molecular typing provides a method for identifying local clusters of gonorrhoea, and could assist in the implementation and evaluation of targeted interventions.

  • Antimicrobial Resistance
  • Gonorrhoea
  • Molecular Typing
  • Molecular Epidemiology

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.


Gonorrhoea is the second most common bacterial sexually transmitted infection (STI) in the UK1 and is associated with significant patient morbidity, including pelvic inflammatory disease in women and epididymitis in men. Gonorrhoeal infection is currently treatable, but concerns have been raised about increasing antibiotic resistance, prompting a recent revision of treatment guidelines in the UK.2

Between 1993 and 2004, the number of episodes of gonorrhoea treated in sexual health clinics in Wales increased steadily to a peak of 23 episodes of uncomplicated gonorrhoea per 100 000 population. Two-thirds of these cases were in the more populous South East of Wales.3

In order to understand better the factors affecting the transmission of gonorrhoea in South East Wales, we characterised gonorrhoea isolates from cases presenting in primary care and Genitourinary medicine (GUM) clinics in the region using Neisseria gonorrhoeae Multi Antigen Sequence Typing (NG-MAST). NG-MAST is a highly discriminatory sequence based typing method4 which examines sequence variation in two hypervariable genes (tpbB and porB) and has been used successfully to investigate sexual networks in Sheffield,5 London,6 Scotland,7 Sweden8 and South Africa,9 together with antimicrobial susceptibility profiles.

Materials and methods


Ethical approval for the study was obtained from the North West Multi-Centre Research Ethics Committee as an amendment of the approval obtained for the Gonococcal Resistance to Antimicrobials Surveillance Programme (GRASP) MREC 00/8/24.


Between May 2005 and September 2006, 540 consecutive gonococcal isolates were collected at three microbiology laboratories in South East Wales. After duplicate isolates representing a single episode of infection were removed (see data management and statistical analysis section below), a total of 511 gonococcal isolates/episodes of infection from 502 patients (figure 1) were collected from the University Hospital of Wales, Cardiff (n=274), Royal Gwent Hospital, Newport (n=176) and Royal Glamorgan Hospital, Llantrisant (n=61). Isolates had been retrieved from patients attending GUM clinics (n=393), general practice (n=76) and other healthcare settings (n=33), while this was unknown for nine cases. The gonococcal isolates were sent to the Health Protection Agency in London and were retrieved by 18–24 h incubation at 36°C in 5% CO2 on GC agar base (Becton Dickinson, Oxford, UK) containing 1% IsoVitaleX, confirmed to be N gonorrhoeae and then stored in glycerol broth at −80°C.

Figure 1

Flow diagram illustrating numbers of Neisseria gonorrhoeae isolates and patients in the study.

Molecular typing

Suspensions (equivalent to a McFarland's standard of 1.0) of retrieved bacterial cultures aged 18–24 h were prepared in phosphate buffered saline, boiled for 10 min and centrifuged for 5 min at 10 000 rpm. The supernatant, containing bacterial DNA, was stored at −20°C for subsequent use in PCR assays.

NG-MAST was performed on all DNA extracts as previously described.4 Sequence data were analysed and edited in BioNumerics V.6.1 (Applied Maths, St-Martens-Latem, Belgium) software. Allele numbers for the trimmed porB (490 bp) and tbpB (390 bp) sequences and sequence types (STs) were assigned via the online NG-MAST database ( Sequence similarity of the STs (concatenated porB and tpbB sequences) from clusters of five or more patients was performed in BioNumerics using Neighbor-Joining multiple alignment.

Susceptibility testing

Susceptibility testing was performed on all retrieved isolates by agar dilution as described previously.10 Minimum inhibitory concentrations (MICs) were determined for seven antimicrobials using the following concentration ranges: ciprofloxacin (0.002–32 mg/l), penicillin (0.03–4 mg/l), tetracycline (1–32 mg/l), ceftriaxone (0.002–0.12 mg/l), cefixime (0.002–0.25 mg/l), azithromycin (0.03–4 mg/l) and spectinomycin (2–64 mg/l). Additionally, β-lactamase activity was determined by the chromogenic indicator Nitrocefin according to manufacturer's instructions (Oxoid, Basingstoke, UK). Categories of resistance were determined using the breakpoints defined in GRASP11 and are listed in the footnote of table 1.

Table 1

Characteristics of 288 isolates and patients for clusters of sequence types (STs) represented by five patients or more

Patient information

Epidemiological data were collected from the notes of patients attending the GUM clinics via a standard questionnaire. These included: gender, site of infection, date of birth, ethnicity, sexual orientation, previous gonococcal infection, symptom presence, concurrent STIs, HIV status and sexual contacts. A more limited data set (gender, date of birth and site of infection) was supplied by the referring laboratories with the bacterial isolates collected from patients attending general practice or other healthcare settings.

Data management and statistical analysis

The epidemiological and laboratory data were linked in Microsoft Access using the patient's clinic number. All analyses were based on distinct episodes of infection, which is defined as a single case of gonorrhoea from one patient within a 28-day period. When more than one isolate was received from an individual within 28 days this was treated as a duplicate and removed unless typing results were different from multiple sites. In this case we assumed that this represented different episodes of infection or mixed infection.

Significance tests by simulation (Fisher's exact tests using 10 000 simulations using StatXact V.7.0.0 (Cytel Inc., Cambridge, Massachusetts, USA)) were performed on the entire data set to explore potential associations between ST and sexual orientation, presence of symptoms, susceptibility profile, age group and diagnostic setting. The characteristics of clusters of five or more individuals were described in more detail. Associations between specific STs and susceptibility profile and sexual orientation were explored by Fisher's exact test in STATA V.11 (Texas, USA).


From the 511 isolates collected in this study, 93% (475/511) were typed using NG-MAST and 68% (348/511) were successfully retrieved from the frozen stock to perform susceptibility testing (figure 1). In all, 17% of the gonococcal population displayed resistance to ciprofloxacin and 5.8% and 34.2% displayed resistance to penicillin and tetracycline, respectively. No resistance or decreased susceptibility to azithromycin, cefixime, ceftriaxone or spectinomycin was demonstrated, and 59.5% (207/348) of isolates were fully susceptible to all antimicrobials.

Patient characteristics

Of the 502 patients in this study, 493 had one episode of infection. Of the remaining nine patients, eight had two episodes of infection, six of which had different strains after 28 days and two patients presented with the same strain after 28 days. The remaining one patient had three episodes of infection, two of which were within 28 days, but with two different strains.

Patient demographic data were available for 507/511 isolates, of which 61% (308/507) were from male patients with a median age of 25 years (range 16–71 years) and 39% (199/507) of isolates were from female patients with a median age of 21 years (range 4–60 years). Overall, 57% (286/501) of patients were aged under 25 years, with an equal proportion of men and women. Additional epidemiological data were available for 393 patients who attended a GUM clinic. Sexual orientation was known for 357 patients and 77% (274/357) were heterosexual men (n=171) and women (n=103) and 23% (82/357) were men who have sex with men (MSM). Chlamydia was the most common concurrent STI with co-infection, documented in 40% (104/274) of heterosexuals (62 men, 42 women; 78% under 25 years (80/103)) and in 12% (9/77) of MSM. Where HIV status was known (n=156), 14% (8/59) of MSM and 2% (2/97) of heterosexuals were HIV-positive.

Sequence typing

A total of 93% (475/511) isolates were characterised by NG-MAST, with 163 different STs identified, representing 128 different porB alleles and 65 different tbpB alleles. Overall, 53 clusters (two or more isolates with the same NG-MAST ST) were identified. A total of 21 clusters were submitted from a single laboratory and 32 clusters included isolates submitted from two or more laboratories. There were 110 STs represented by a single isolate. The most frequently observed types (>20 isolates) were: STs 2, 752, 471, 249 and 8 (table 1) accounting for, respectively, 9.5%, 5.7%, 5.3%, 4.8% and 4.6% of all isolates typed (29.9%; 142/475 overall), all of which were fully susceptible to the antimicrobials tested.

Predominant ST clusters

The largest ST clusters, those with five or more isolates, represented 288 cases of gonococcal infection (61% of all typed isolates, 288/475) identified in patients attending either STI clinics or other healthcare settings (table 1). Significant associations between ST and antimicrobial resistance profiles were observed, with 210/219 (96%) isolates from 25 clusters having a consensus antimicrobial resistance profile, with the discrepancies mainly being with MICs close to a breakpoint (resistance category may vary even if the MICs are just one dilution different). While isolates in the majority of STs were generally sensitive to all antimicrobials tested, isolates of STs 495, 1609 and 960 were significantly associated with tetracycline resistance (p values ranging from 0.001 to 0.03) and isolates identified as STs 225, 147, 1634 and 404 were typically resistant to ciprofloxacin and tetracycline (p values ranging from <0.001 to 0.03) while ST 597 was additionally resistant to penicillin (p=0.009) (table 1).

Of the 288 patients infected with a frequently occurring ST, 222 had attended a GUM clinic which collected additional patient information (table 1), including sexual orientation. Statistical analyses of this subset of patients showed a significant association between ST and sexual orientation (p<0.001) (table 1). STs 147, 4, 1634 and 64 were significantly associated with MSM patients (p values ranging from <0.001 to 0.001) while the association between STs 225, 960, 210, 597 and 359 and MSMs was of borderline significance (p values ranging from 0.01 to 0.04). The ST 597 cluster contained three individuals who were diagnosed with HIV at the same time as their gonococcal infection. The phylogram (figure 2) shows further clustering among STs, sexual orientation and susceptibility profile of STs represented by five patients or more. There is clear clustering of STs, susceptibility profile and sexual orientation in groups A, C, E and F. All groups, other than group B, show clustering of susceptibility profile.

Figure 2

Number of isolates in each sequence type (ST) cluster (≥5) and associated sexual orientation. (Note: Sexual orientation of clusters is defined at 70% of patients reporting sexuality). FS, fully susceptible; Hetero, heterosexual; Mixed, groups composed of both MSM and heterosexuals; MSM, men who have sex with men.

Most of the STs shown to be significantly associated with resistance to at least one antimicrobial agent were also associated with infection in MSM patients, with the exception of STs 495 and 1609. ST 4 was the only genotype associated with MSM patients for which the isolates were sensitive to all antimicrobials tested.

Linking of genotyping and contact tracing data

Extensive contract tracing was not performed but patients were asked to name sexual contacts. Of the patients of known sexual orientation for whom a NG-MAST type was determined, 113 heterosexual patients and 31 MSM patients named a sexual contact. In all, 96 of these named contacts were participants in this study. The concordance of the STs between the sexual contacts was established for heterosexual patients (n=81) and MSM (n=15). Within the MSM group, there were four partner pairs with mutual naming and the same ST, and two instances of three individuals in a transmission chain with the same ST, although one cluster showed additional contacts infected with a different ST. In the heterosexual group there were 25 mutually-named partner pairs infected with the same ST and eight patients who named a sexual partner that did not name them, but both individuals were infected with gonococci of the same ST. There were two patient clusters containing three individuals infected with two STs and one instance of three individuals in a transmission chain infected with the same ST. Four matched partner pairs were infected with two different STs. In one of these pairs, the gonococcal isolation date was more than 1 month from each other.

Of the three sexual contact pairs with non-concordant STs, the porB and tbpb alleles were both different suggesting different strains, except in one patient who had the same tbpb allele (16), but different porB alleles (289 and 959) which differed from each other by two nucleotides. The NG-MAST procedure failed to produce a complete ST for each patient in eight mutually-named partner pairs.


This is the first study to have explored the molecular epidemiology of an STI in Wales and we confirm the existence of common STs across the UK as well as identifying a number of types that were novel to Wales. Discrete sexual networks were identified, the most localised being in young heterosexuals.

The most frequently observed gonococcal STs were 2, 752, 471, 249 and 8, accounting for almost 30% of all isolates typed. STs 2 and 8 were among the most frequently observed types in earlier studies in London in 20004 and 2004,6 and in Scotland from 2004 to 2006,12 suggesting widespread and sustained distribution of these types in the heterosexual population in the UK. STs 471 and 249 were also among the larger ST clusters identified in Scotland from 2004 to 2006 but ST 752 has not been reported previously as a predominant type in any of the studies which have applied NG-MAST to characterise consecutive isolates in a population.4 ,6 ,12 ,13 This may reflect the limited information available on prevalence of STs in geographical regions of the UK, or the predominance of ST 752 may represent a specific characteristic of the gonococcal population within sexual networks in Wales.

While the unexpectedly low retrieval rate of 68% is a limitation of the current study, the observed associations between ST and antimicrobial resistance nevertheless demonstrate the potential utility of NG-MAST as an adjunct to gonococcal nucleic acid amplification tests to predict the likely susceptibility pattern to aid appropriate patient management. For example, the observed association between ciprofloxacin resistance and STs 225, 147, 1634, 597 and 404 demonstrates that use of fluoroquinolones to treat these types would be highly inappropriate. Associations between ciprofloxacin resistance and STs 14712 ,14 ,20 and 22512 ,14 ,17 have been documented previously in other European studies, but to our knowledge this has not been reported before for STs 1634, 597 or 404. All of these STs were predominantly in MSM, which is consistent with the observation in GRASP11 that the burden of ciprofloxacin resistance is within this patient group. This study has provided further insight into the epidemiology of ciprofloxacin resistance by showing it was confined to a limited number of types. The third generation cephalosporin, ceftriaxone, is now recommended for first-line therapy rather than ciprofloxacin.2 Isolates exhibiting decreased susceptibility to cefixime that were predominantly ST 1407 or closely related types have emerged recently in the UK.10 In the current study there was one example of ST 1407, recovered from a female in Cardiff in 2005, which while not showing decreased susceptibility, did have a raised MIC to cefixime. Thus, NG-MAST typing may prove useful in predicting which infections have a greater risk of treatment failure in the future. The potential application of NG-MAST to predict antimicrobial susceptibility in gonococci has been proposed previously12 and is attractive as it could provide additional information for the management of patients in countries where culture is not performed routinely, or where infection is at a site where the sensitivity of culture is low. However, it should be appreciated that antimicrobial resistance can emerge rapidly in N gonorrhoeae and so this approach would require extensive ongoing surveillance of circulating STs in relation to antimicrobial susceptibility in that geographical region to ensure accurate prediction.

As in previous studies, the current study demonstrated a clear association between ST and sexual orientation. STs 2, 8, 51 and 25 in London4 ,6 ,15 and ST 249 in Edinburgh, Scotland,16 have been noted previously to predominate in heterosexual patients, while the associations between STs 147, 225, 64, 210 and 359 and MSM were also reported in patients in London.4 ,6 ,15 A study of the gonococcal population in patients attending an STI clinic in Edinburgh also showed STs 4, 64, 210 and 225 were MSM-associated.16 STs 147, 225 and 359 have been reported in MSM in other European countries,17–19 although the Scottish study and another conducted in Sweden suggested that ST 147 can be observed in a significant proportion of heterosexual patients also.16 ,20 To our knowledge, the association between MSM and STs 1634, 960 and 597 has not been reported previously. It is evident that particular STs are widely disseminated in the UK in different population groups. The persistence of these types in MSM and heterosexual groups in the current and previous studies over a 3-year period demonstrates their stability and the continued lack of bridging between these networks, for example, via bisexual males.

As shown previously5 ,16 there was high concordance (83%) between NG-MAST STs and named sexual contacts demonstrating again the application of NG-MAST as a confirmatory tool with contact tracing. However, the use of NG-MAST alone in social network analysis does have limitations as STI networks are not closed and inference on which network or cluster an individual belongs to should be taken with great care. Since 2004, the number of episodes of gonorrhoea reported has declined slightly. However, in 2009 there was an increase to 16 per 100 000 population. This study has shown that NG-MAST has the potential to be a powerful public health tool. Identification and characterisation of localised sexual networks provide the opportunity for targeted interventions, including intensive contact tracing, and it appears likely that this would be most successful in networks of young heterosexuals. Public Health Wales is currently piloting dual chlamydia and gonorrhoea testing in low prevalence community settings. Typing would provide useful information about transmission in low prevalence settings, and assist in quantifying the level of ‘bridging’ between ‘high-risk’ populations accessing services in hospital and those, often from more rural areas, accessing community settings.

Key messages

  • This is the first molecular epidemiology study of a sexually transmitted infection in Wales.

  • Common Neisseria gonorrhoeae Multi Antigen Sequence Typing sequence types were identified from across the UK, as well as a number of types that were novel to Wales.

  • Discrete sexual networks were identified, the most localised being in young heterosexuals.

  • Molecular typing provides a method for identifying local clusters of gonorrhoea, and could assist in the implementation and evaluation of targeted interventions.


We thank all the staff at the collaborating microbiology laboratories and GUM clinics, in particular Dr Robert Das who was involved in the study at the Royal Gwent Hospital, Newport. A special thank you to the members of STBRL who performed the susceptibility testing and assisted with the molecular typing. We would also like to express thanks to Tom Nichols for performing the statistical analysis and Dr Mark Temple for providing advice on the feasibility of sexual network analysis in this sample.



  • Contributors MJC and SAC contributed to the laboratory work, data analysis and manuscript preparation. DRT contributed to the design of the study, collated all the epidemiological data and assisted with the data analysis and manuscript preparation. IM contributed to the design of the study and laboratory work. CAI contributed to the design of the study, supervised the project and contributed to the manuscript preparation. ANA, HB, IH, EK and DW provided the specimens and epidemiological information and contributed to the design of the study. All authors reviewed the final draft of the manuscript.

  • Funding The work was supported by internal funding.

  • Competing interests None.

  • Ethics approval North West Multi-Centre Research Ethics Committee.

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