To analyse trends in antimicrobial resistance (AMR) in Neisseria gonorrhoeae (GC) isolated in Australia between 1997 and 2006 and to identify factors influencing emergence and spread of AMR in GC.
AMR data were generated in reference laboratories in each state and territory of Australia using the methods of the Australian Gonococcal Surveillance Programme from a comprehensive sample of GC. Trends in the proportion of strains resistant to penicillin, ciprofloxacin, spectinomycin and ceftriaxone or with high-level tetracycline resistance (TRNG) were determined from aggregated national data and were also disaggregated by region. Further analyses of additional AMR, demographic, transmission and antibiotic use data were also performed.
More than 36 000 GC were examined. Significant increases in resistance to penicillin and ciprofloxacin and in TRNG occurred in national data and in urban populations. Approximately half of the GC tested in larger urban centres were penicillin and/or ciprofloxacin resistant by 2006. These high rates of resistance arose despite low (penicillin) or absent (ciprofloxacin) exposure. In contrast, in rural and remote areas with very high disease rates and high rates of penicillin use, <5% of GC tested were penicillin (or quinolone) resistant. No spectinomycin-resistant GC were detected. Low numbers of GC with raised MICs of ceftriaxone were present in urban centres each year from 2001 onwards.
Significant increases in AMR in GC occurred in parts of Australia in the 10 years to 2006. The data suggest that the AMR seen in GC in urban populations were the result of their repeated importation into Australia and ultimate introduction into established sexual networks rather than originating de novo or as a result of selection by antibiotic use or misuse.
Surveillance for antimicrobial resistance (AMR) is used to guide therapy in the individual patient, monitor AMR trends, track the emergence and spread of resistance or to allow advocacy for rational use of antibiotics.1,2 Additionally, in diseases of public health importance such as gonorrhoea, AMR surveillance helps establish and maintain the efficacy of standard treatment regimens to assist in disease control.1–3 Control of gonorrhoea is complex and difficult, requiring an integrated and multidisciplinary approach that includes effective antibiotic treatment.4 This is defined as those single dose regimens, provided at first diagnosis, that result in a cure in 95% or more of cases.4–6 Their formulation thus requires knowledge of AMR patterns in gonococci obtained from surveillance.3–5,7 Programmes for reliable AMR surveillance for Neisseria gonorrhoeae need to accommodate the fastidious gonococcal growth requirements, the propensity of the organism to rapidly acquire resistance and to conform to recognized surveillance standards.1–5
The Australian Gonococcal Surveillance Programme (AGSP) is an established collaborative programme of AMR surveillance conducted in each jurisdiction in Australia.8 It differs from some equivalent programmes9–11 in that it is comprehensive (as opposed to sentinel), continuous (rather than intermittent) and actively integrates AMR analysis with selected clinical data. The origins, procedures and early outcomes of the programme have been published.8,12,13 In addition, the AGSP has provided, in quarterly reports published continuously from 1981 in public health bulletins, timely data on the susceptibility of gonococci to those antibiotics relevant to treatment of gonorrhoea in Australia together with detailed annual analyses from 1996 onwards (available at a dedicated URL).14 Audits of the AGSP indicate that it meets or exceeds World Health Organization and CDC performance criteria for surveillance programmes of this type.15
This report provides further analyses of trends in AMR in gonococci isolated in Australia over the 10 years 1997–2006. Gonorrhoea in Australia has some distinctive features. A largely urban population living mainly in coastal areas has a disease rate and distribution similar to that reported in other Western-style industrialized countries, and with decreasing options for treatment because of increasing AMR. However in non-urban settings in ‘outback’ locations where older treatments, including the penicillins, still remain highly effective, the disease incidence is much higher, although this high rate yields only relatively small numbers of cases due to a low population density. There are also differences in rural and urban disease patterns and some unusual disease manifestations occur.16,17 This paper includes a consideration of these elements to discuss factors influencing the emergence and spread of AMR in gonococci in Australia.
Materials and methods
The structure, function and methods employed by the AGSP have been described previously.8,14,18 Briefly, gonococcal isolates from a comprehensive network of private and public sector laboratories are forwarded to reference centres in each jurisdiction for examination using standardized methodologies. The AGSP determined the MICs of penicillin, ceftriaxone, spectinomycin and ciprofloxacin using agar plate dilution methods with 104 cfu per spot inoculated onto Sensitest agar (Oxoid, Basingstoke, UK) supplemented with 8% saponin-lysed horse blood and appropriate dilutions of the relevant antibiotic (obtained from manufacturers) and incubated in 5% CO2 in air for 24 h. Quality control (QC) strains used in each laboratory, originally WHO A—E, were progressively updated18 through collaboration with other international centres. An external programme-specific quality assurance process (EQAS) was conducted by a co-ordinating centre.19 Comparability of data was obtained by the uniformity of methods and the QC and EQAS programmes. Dilution ranges and resistance criteria were described initially for penicillin8 and with subsequent additions as newer agents were introduced.18 For penicillin (CMRP), ceftriaxone and ciprofloxacin, chromosomally mediated resistance was defined as an MIC ≥ 1 mg/L and decreased susceptibility in the range 0.06–0.5 mg/L. Spectinomycin resistance was defined as an MIC ≥ 64 mg/L and high-level tetracycline resistance (TRNG) as an MIC ≥ 16 mg/L.20 Penicillinase-producing N. gonorrhoeae (PPNG) comprised a separate category of penicillin resistance.
Data from each participating centre were forwarded to the co-ordinating centre for collation, analysis and reporting. Data recorded, as well as the number and proportion of isolates in the resistance categories defined above, included the date and anatomical site of infection, the sex of the patient and post code of residence, and for certain resistant isolates, details of where the infection was acquired.15 A combined analysis of data generated by the AGSP over the period 1 January 1997 to 31 December 2006 provided the basis of this study.14 Significant differences in resistance frequencies, categorized as those at a value of P ≤ 0.05, were determined by defined protocols1 and as described previously.21 Previously published material on gonococcal diseases and their distribution and patterns in Australia, antibiotic usage data and comparable resistance rates in other relevant community-acquired infections was reviewed and relevant information analysed in conjunction with the AMR data.
Numbers and distribution of isolates tested
The AGSP examined 36 258 gonococci in the 10 year period 1 January 1997 to 31 December 2006. Numbers examined in each year ranged between 2817 in 1997 and 3886 in 2005. These numbers represent about two-thirds of all laboratory-based notifications of gonorrhoea annually in Australia.15 Slightly more than half of these isolates came from the states of New South Wales (30%) and Victoria (25%), with the jurisdictions of Western Australia, Queensland, South Australia and the Northern Territory contributing between 5% and 15% of all isolates each. There were only small numbers of gonococci from Tasmania and the Australian Capital Territory in any year. The numbers of isolates by jurisdiction for each year vary slightly14 but by way of illustration those for 2006 were: 1198 and 951 for the states of New South Wales and Victoria (the two most populous jurisdictions); 565, 397 and 244 for Queensland, Western Australia and South Australia, respectively (states with the majority of their population concentrated close to their respective capitals, but each of a large area and with high disease rates in rural populations); 549 from the Northern Territory (again a large area with a low population density but also with high disease rates); 14 from the island state of Tasmania; and 19 from the Australian Capital Territory. Although there have been amalgamations of both public and private sector laboratories over time, the patterns and ultimate source of referrals have altered but little over this period of the programme.
Aggregated national trend data over 10 years to 2006 for resistance to the penicillins and ciprofloxacin and for high-level TRNG are shown in Figure 1. There were several significant (P < 0.05) increases and decreases in the proportion of isolates resistant to penicillin and those gonococci non-susceptible to ciprofloxacin from 1997 to 2003 after which there was a continuing and significant increase (as above) in resistance to both antibiotics. Nationally by 2006, over one-third of isolates was resistant to at least one of these agents. The proportion of isolates displaying high-level tetracycline resistance increased more regularly from 5% in 1997 to 12% in 2006. No isolates resistant to spectinomycin were detected in the 10 year period. A total of 134 isolates with ceftriaxone MICs in the less susceptible range (0.06–0.5 mg/L) were detected in urban centres from 2001 to 2006; 108 from New South Wales, 14 from Queensland and 7 from Victoria with small numbers from South Australia, Western Australia and the Northern Territory. Most of these isolates were also multiresistant (to penicillins and quinolones also) and were isolated mainly from overseas travellers or their contacts, but also from domestic transmission.
Disaggregated trend data showing the proportion of PPNG and chromosomally mediated penicillin resistance to penicillin (CMRP) in New South Wales and Victoria and for all penicillin resistance in the Northern Territory are shown in Figure 2, and for isolates less susceptible and resistant to ciprofloxacin in New South Wales and Victoria and all non-quinolone susceptible gonococci in the Northern Territory in Figure 3. These figures contrast resistance patterns in the largely urban samples obtained from New South Wales and Victoria with those seen in the essentially non-urban population of the Northern Territory.
In New South Wales and Victoria, there were relatively close parallels between trend data for the proportions of PPNG and CMRP present over 10 years (Figure 2). The proportion of PPNG in both centres ranged between 6% and 14% but a much greater volatility was observed for CMRP. Approximately 50% of gonococci tested in each of these centres were penicillin resistant by any mechanism by 2006. In contrast in the Northern Territory, the proportion of penicillin-resistant gonococci (by any mechanism) was substantially lower, <5%. Most of this resistance in the Northern Territory was due to PPNG encountered in travellers or seamen in the coastal capital, Darwin. Very few CMRP were encountered and none in remoter areas of the Territory.
A similar pattern is seen in Figure 3 where the proportion of gonococci less susceptible or else resistant to ciprofloxacin also approximated closely in New South Wales and Victoria. Again, the trend line for gonococci non-susceptible to the quinolones at any MIC level in the Northern Territory was quite dissimilar to that seen in New South Wales and Victoria, and the resistant isolates were almost always from disease acquired external to Australia. The proportion of isolates less susceptible to the quinolones declined to very low levels from 2001 onwards in New South Wales and Victoria, whereas the proportion of resistant strains increased progressively and particularly from 2003 onwards, ultimately approximating 45% of all isolates and mainly from sustained domestic transmission. The ciprofloxacin MICs for the majority of the resistant isolates were between 8 and 64 mg/L in this period.14 In contrast, the proportion of quinolone-resistant N. gonorrhoeae (QRNG) seen in the Northern Territory remained within a narrow range for the entire period and represented imported isolates.
Data from 2006 on resistance to penicillins and ciprofloxacin and TRNG in South Australia and Western Australia, and also disaggregated within these jurisdictions, are shown in Table 1 and compared with national averages. In this example from jurisdictions comprising two of the larger geographic areas in Australia, there is again a significant difference (P < 0.05) between the higher levels of resistance to penicillins and quinolones in metropolitan areas and equivalent data from remote parts of these states.
|WA||PEN all R||19.2||32.6||3.6|
|SA||PEN all R||17.3||27||0|
|Australia||PEN all R||34|
|WA||PEN all R||19.2||32.6||3.6|
|SA||PEN all R||17.3||27||0|
|Australia||PEN all R||34|
Previously published data13,14,22,23 have shown that in large urban settings in Australia, a substantial proportion of gonorrhoea is concentrated in homosexually active men, with smaller contributions to disease numbers from infected travellers and their contacts and the contacts of commercial sex workers. For example, disease rates in inner Sydney health areas, with a high concentration of homosexually active men, were of the order of 60–80/100 000 population, whereas rates in other nearby urban health areas were <10/100 000 population.24 This disease pattern contrasted with the essentially heterosexual disease transmission patterns in non-urban and remote areas where there is little direct or indirect overseas contact but where disease rates may exceed 1000/100 000. These different patterns of disease were reflected in the high male-to-female disease ratio (more than 10:1) and the high proportion of male rectal and pharyngeal isolates examined in urban centres.14 Again this pattern contrasted with a male-to-female disease ratio of 2:1 in regional centres where almost all isolates were from genital tract samples.14
The nature of gonococcal disease in urban and non-urban settings also differed in some important aspects. Higher rates of disseminated gonococcal infection were seen in rural and remote Australia16 and there were recurrent epidemics of gonococcal conjunctivitis in aboriginal toddlers.17 Published and subsequent unpublished data showed the presence of highly dissimilar gonococcal serogroups circulating in Sydney and the Northern Territory.16,25 IA serogroups were significantly more prevalent in the Northern Territory, whereas IB serogroups comprised more than 85% of all isolates in Sydney.
In keeping with studies elsewhere,6,7 AMR in N. gonorrhoeae isolated in Australia has increased significantly in recent years, especially to the penicillin and quinolone groups of antibiotics. However, these increases have been uneven and occurred much earlier than equivalent changes in Europe or North America.6,7,9 In remote centres of Australia where oral penicillins still remain the standard treatment regimen, resistance rates were below the critical 5% level for regimen change in the 10 years of this study. In contrast, in a number of the larger urban centres, high rates of penicillin and quinolone resistance were documented prior to the beginning of this study26,27 and this necessitated a shift to treatments based on the injectable cephalosporin, ceftriaxone. During the period of this study, further significant increases in the proportion of gonococci resistant to penicillin and ciprofloxacin occurred in more of the larger urban centres, but after some years discontinuation of these antibiotics as treatment options for gonorrhoea.
In the latter period of this study (2004–06), rapidly increasing resistance to penicillins occurred in urban centres at a time when use of these agents had been discontinued for treatment of gonorrhoea (Figures 1 and 2), and when the general usage of β-lactam antibiotics had also decreased nationally.28 Further, quinolone resistance in N. gonorrhoeae increased to unprecedented levels (in terms of the proportion of resistant strains encountered—for example ∼45% in New South Wales—and the progressively higher MICs recorded for these resistant strains—up to 64 mg/L),14 from 2003 to 2006. Prior to and during this period, general use of quinolone antibiotics for any purpose nationally remained very low in human and animal medicine, mainly because of restrictions on government subsidies for quinolone treatments. Over the same period, results of AMR surveillance of quinolone resistance in other community-acquired pathogens such as Campylobacter sp. from humans and animals29,30 and Escherichia coli from urinary tract infections in Australia31 revealed very low levels of resistance to quinolones in these ‘indicator’ organisms that in other settings have shown significant increases in quinolone resistance following direct or indirect quinolone exposure. However, these community-acquired organisms have markedly less efficient transmission mechanisms than gonococci.
Single-dose programmatic treatment of gonorrhoea helps increase compliance and contain AMR in GC4 although some contend this practice may actually select for AMR.32 Australian data documenting increases in QRNG despite low quinolone usage do not support this contention. However, certain cities in Australia are susceptible to the repeated importation of antibiotic-resistant gonococci highly prevalent in nearby regions.33 The increase in ciprofloxacin resistance over time in Sydney, New South Wales, has been serially documented.2,26,27,34 Over a prolonged period to 1991, continuing importation of diverse subtypes of gonococci with raised ciprofloxacin MICs occurred, but only in low proportions.34 A series of periodic rapid increases in the proportion of QRNG with progressively higher MICs to ciprofloxacin then followed. The QRNG involved comprised a limited number of resistant subtypes that eventually became established in local transmission chains26,27 resulting in their sustained domestic transmission. Although similar events were subsequently described for other overseas centres,7,35 the current pattern of QRNG isolation in Darwin, Northern Territory, corresponds to the much earlier events in Sydney. In remote communities in Australia, there is very little resistance despite highly efficient disease transmission and use of penicillins over many decades. Here, there is little contact with overseas travellers or bridging between rural and urban populations. These communities also have documented clusters of disease uncommonly seen in urban centres16,17 and substantially different gonococcal subtypes.
This combination of an analysis of AMR, demographic, transmission and antibiotic use data together suggests the spread of resistant gonococci occurred in Australia when imported organisms, incidentally but already antibiotic resistant, became established in certain patient subgroups, but for reasons other than antibiotic selection pressures36 and when the relevant antibiotics were no longer in use. Control of AMR in gonococci is difficult because of its capacity for genetic recombination and high transmissibility that includes a significant capacity for regional and international spread.4,37 These factors, when coupled with a propensity to develop AMR through over use and misuse of antibiotics in settings where disease control is also problematic, suggest that broader, global initiatives that integrate both disease and AMR control initiatives are urgently needed.37
Even if the above measures are in place, they will probably only delay rather than prevent emergence of further AMR in gonococci. The Australian data included small numbers of gonococci with decreased susceptibility to ceftriaxone. Other studies show that gonococci with these characteristics are becoming more widespread.35,38,39 The mechanisms of resistance involve altered penA40,41 genes but are not as yet fully elucidated and thus far this resistance appears to impact more on the clinical efficacy of oral third-generation cephalosporins than on ceftriaxone.40–43 If this form of resistance evolves further, efficient methods for its detection and control will need to be in place.44 Continuing surveillance of AMR in N. gonorrhoeae, and at an enhanced level, would thus seem to be prudent despite the considerable technical and logistical difficulties associated with this approach.
The Department of Health and Ageing of the Commonwealth Government of Australia provides financial assistance to the National Neisseria Network of Australia of which the AGSP is a component. There were no other funding sources.
John Bates, Vicki Hicks (Queensland Health Scientific Services, Coopers Plains, Queensland); Tiffany Hogan, Sanghamitra Ray (Department of Microbiology, The Prince of Wales Hospital, Randwick, New South Wales); Julia Griffith, Jocelyn Hibberd, Mark Veitch, Geoff Hogg (The Microbiological Diagnostic Unit (PHL), Department of Microbiology and Immunology, University of Melbourne, Parkville, Victoria); Ann Weaver (Infectious Diseases Laboratories, Institute of Medical and Veterinary Science, Adelaide, South Australia); Julie Pearson, Hui-Leen Tan (Department of Microbiology and Infectious Diseases, PathWest Laboratory Medicine, Royal Perth Hospital, Western Australia); Mark Gardam, Alistair Macgregor (Department of Microbiology and Infectious Diseases, Royal Hobart Hospital, Hobart, Tasmania); Gary Lum and Microbiology Staff (Microbiology Laboratory, Royal Darwin Hospital, Casuarina, Northern Territory); and Susan Bradbury, Peter Collignon (Microbiology Department, Canberra Hospital, Woden, Australian Capital Territory).
Many other colleagues contributed to various phases of the development of the AGSP, and we thank particularly J. R. L. Forsyth, Margaret Peel and Greg Handke.
None to declare.
- gonococcal infection
- drug resistance, microbial
- neisseria gonorrhoeae
- tetracycline resistance
- urban population
- surveillance program
- malnutrition-inflammation-cachexia syndrome
- remote environment