Article Text

Download PDFPDF

Prevalence, incidence and risk factors for pharyngeal gonorrhoea in a community-based HIV-negative cohort of homosexual men in Sydney, Australia
  1. David J Templeton1,2,
  2. Fengyi Jin1,
  3. Leon P McNally3,
  4. John C G Imrie4,
  5. Garrett P Prestage1,
  6. Basil Donovan1,5,
  7. Philip H Cunningham3,
  8. John M Kaldor1,
  9. Susan Kippax4,
  10. Andrew E Grulich1
  1. 1National Centre in HIV Epidemiology and Clinical Research, The University of New South Wales, Sydney, Australia
  2. 2RPA Sexual Health, Royal Prince Alfred Hospital, Sydney, Australia
  3. 3SydPath, St Vincent's Hospital, Sydney, Australia
  4. 4National Centre in HIV Social Research, The University of New South Wales, Sydney, Australia
  5. 5Sydney Sexual Health Centre, Sydney Hospital, Sydney, Australia
  1. Correspondence to Dr David J Templeton, National Centre in HIV Epidemiology and Clinical Research, The University of New South Wales, Level 2, 376 Victoria Street, Sydney NSW 2010, Australia; dtempleton{at}nchecr.unsw.edu.au

Abstract

Background Pharyngeal gonorrhoea is common in homosexual men and may be important in maintaining community prevalence of anogenital infections.

Methods From 2003, all participants in the Health in Men cohort of HIV-negative homosexual men in Sydney were offered annual pharyngeal gonorrhoea screening by BD ProbeTec nucleic acid amplification (NAAT) assay with supplementary porA testing. Participants self-reported diagnoses of pharyngeal gonorrhoea made elsewhere between interviews. Detailed sexual behavioural data were collected 6-monthly.

Results Among 1427 participants enrolled, 65 study-visit-diagnosed pharyngeal gonorrhoea infections were identified (incidence 1.51 per 100 person-years, 95% CI 1.19 to 1.93) of which seven infections were identified on baseline testing (prevalence 0.57%, 95% CI 0.23 to 1.17%). Almost 85% of study-visit-diagnosed pharyngeal infections occurred without concurrent anogenital gonorrhoea. The combined incidence of study-visit-diagnosed and self-reported pharyngeal gonorrhoea (n=193) was 4.45 per 100 person-years (95% CI 3.86 to 5.12). On multivariate analysis, incident infection was associated with younger age (p-trend=0.001), higher number of male partners (p-trend=0.002) and reported contact with gonorrhoea (p<0.001). Insertive oro-anal sex (‘rimming’) was the only sexual behaviour independently associated with incident pharyngeal gonorrhoea (p-trend=0.044).

Conclusions The majority of pharyngeal gonorrhoea occurred without evidence of concurrent anogenital infection, and the high incidence-to-prevalence ratio suggests frequent spontaneous resolution of NAAT-detected infection. The association of pharyngeal gonorrhoea with oro-anal sex indicates that a broader range of sexual practices are likely to be involved in transmission of gonorrhoea to the pharynx than previously acknowledged. Screening the pharynx of sexually active homosexual men could play a role in reducing the prevalence of anogenital Neisseria gonorrhoeae.

  • Neisseria gonorrhoeae
  • pharynx
  • homosexuality
  • male
  • prospective studies
  • epidemiology
  • gay men
  • neisseria gonorrhoea
  • pharynx
View Full Text

Statistics from Altmetric.com

Introduction

Neisseria gonorrhoeae infection of the pharynx has been recognised for decades1 2 and is commonly identified among men who have sex with men (MSM) in community3 4 and clinic-based5–7 settings. In contrast to older studies,8–10 recent data suggest that the pharynx is the most common site of gonococcal infection in MSM.3 5 6 However, there are no prospective community-based data on the proportion of pharyngeal gonorrhoea which occurs independent of anogenital infection.

The high prevalence of gonorrhoea among MSM is not surprising, given that the gonococcus is more efficiently transmitted via fellatio than cunnilingus,9 oral sexual behaviours are substantially more common among homosexual than heterosexual men,11 12 and behaviour change since the mid-1980s has led to reduced rates of unprotected anal, but not oral, intercourse.13 We have previously described an association of urethral and anal gonorrhoea with a range of oro-genital and oro-anal sexual behaviours14 in the Health in Men (HIM) cohort. These and other observational data15 suggest that the pharynx is a likely source of anogenital gonorrhoea in MSM.

The single published longitudinal study of risk factors for pharyngeal gonorrhoea in MSM assessed only penile–oral contact as a sexual behavioural predictor of infection.4 However, MSM practice a range of other oral sexual behaviours, none of which have been prospectively investigated as predictors of pharyngeal gonorrhoea. Here, we describe the prevalence, incidence and predictors of pharyngeal N gonorrhoeae among the community-based Health in Men (HIM) cohort of homosexual men in Sydney, Australia.

Materials and methods

Participants

Study participants were men who were recruited from a range of community-based sources between June 2001 and December 2004, and followed until June 2007. Details of recruitment sources and methods of the HIM study have been described elsewhere.16 Briefly, men eligible for participation in the HIM study met the following criteria: (1) they reported having sex with men in the 5 years prior to enrolment, (2) they lived in, and/or participated in, the gay community of Sydney and (3) they tested negative for HIV at baseline. Signed, informed consent was obtained from all participants. Ethics approval was granted by the University of New South Wales.

Data collection

All participants underwent annual face-to-face interviews, with 6-monthly telephone interviews between these visits. Participants who were diagnosed as having HIV infection during the study period were excluded from that time point onwards. At baseline and annual face-to-face interviews, participants were asked to report if they had been diagnosed as having gonorrhoea in the throat or had sexual contact with gonorrhoea in the last 12 months. A history of recent sore throat was sought at annual visits. Oral sexual practices were reported by participants every 6 months and are defined in the footnotes of table 1.

Table 1

Univariate analysis of sexual behavioural risk factors with casual partners in last 6 months for pharyngeal N gonorrhoeae infection in the HIV-negative HIM cohort among men who reported casual partner(s)

Laboratory studies

Nucleic acid amplification tests (NAAT) for N gonorrhoeae and Chlamydia trachomatis were introduced to the HIM study in January 2003. Anal swabs and urine testing were also offered to participants as previously described.14 The HIM study nurse performed a pharyngeal swab on consenting participants by wiping the tonsils, tonsillar crypts and posterior pharynx at the time of the annual face-to-face interview. Swabs were tested by strand displacement amplification (SDA) using the BD ProbeTec assay (BD Diagnostics, Sparks, Maryland) as per the manufacturer's instructions. All positive BD ProbeTec N gonorrhoeae pharyngeal and anal samples underwent supplemental testing by an assay targeting the gonococcal porA pseudogene. This supplemental assay has previously been shown to have excellent sensitivity and specificity for non-genital gonorrhoea diagnosis in a population of predominantly homosexual Australian men.17 The laboratory methods of this supplemental testing have been described in detail previously.18 Men who tested positive were referred to their usual doctor for treatment, usually with a single intramuscular dose of 250 mg ceftriaxone.

Incidence definition

Incidence of self-reported diagnosis of pharyngeal N gonorrhoeae

Participants who reported a diagnosis of pharyngeal gonorrhoea in the last 12 months were treated as self-reported incident cases (‘interval diagnoses’). Australian guidelines recommend the use of culture to screen for pharyngeal gonorrhoea in asymptomatic MSM.19 To incorporate baseline information, because participants reported a 12-month history of diagnosis, one person-year (PY) was allocated for every participant who responded to the question at baseline interview, whether or not they reported infection. An identical definition was used in recently published analyses of urethral and anal infections14: in that analysis, inclusion of the baseline data allowed for much greater study power but did not change point estimates of effect.

Incidence of NAAT confirmed diagnosis of N gonorrhoeae

Participants who tested positive to pharyngeal N gonorrhoeae on both ProbeTec and supplemental porA assays were treated as NAAT confirmed incident cases (‘study visit diagnoses’). In order to incorporate baseline data for incidence analyses, one PY was allocated for participants' initial tests, as described above.

Combined incidence of interval and NAAT diagnosis of N gonorrhoeae

To be able to estimate more comprehensively the incidence of pharyngeal gonorrhoea, interval and study visit diagnoses of pharyngeal N gonorrhoeae were combined. Interviewers were instructed to distinguish study visit diagnoses from interval diagnoses, so diagnoses made at last face-to-face interview would not be reported as interval diagnoses at the current interview. Since the NAAT testing was added to the study in January 2003, this calculation was only applicable from 2003 forward. Participants who answered the self-report question but declined testing were included in the denominator for this combined incidence.

Statistical analysis

Statistical analyses were performed using STATA 10.0 (STATA Corporation). The exact binomial method was used to calculate 95% CIs for prevalence and incidence values. For the calculation of incidence, the date of infection was assumed to be the midpoint between the interview when participants had an interval or study visit diagnosis and the last face-to-face interview. For cases that were diagnosed at the baseline interview, the date of infection was assumed to be 6 months prior. Multiple failures in the subsequent interview(s) were allowed for the calculation of incidence. Total PY were calculated as the time from study entry to the last face-to-face interview in the HIM study. Risk factors were examined individually for interval and study visit diagnoses, but as results were very similar (data not shown), risk factors for the occurrence of pharyngeal gonorrhoea are presented based on the combined incidence estimate. Cox regression was used to identify risk factors associated with incident infections. Crude and adjusted analyses were performed to identify factors associated with incident pharyngeal gonorrhoea infection. Hazard ratios (HRs) and their corresponding 95% CIs were calculated for these associations. Multivariate Cox regression models were developed to determine risk factors which were independently associated with incident infections. For risk-factor analyses, mixed models were used that allowed for repeated measures in the same individuals. These models, which allow for within- and between-subject variability, provide robust variance estimates, and therefore appropriate p values and CIs.20 Two separate multivariate models were developed: model 1 for the association with demographics and sexual behaviour with regular partner(s); and model 2 for the association with demographics and sexual behaviour with casual partners, using forward stepwise regression. The two models included variables that had univariate HRs with 95% CIs excluding 1.0. They were entered by the order of the magnitude of Wald χ2 value. The association with a broad range of sexual behaviours with regular and casual partners that could plausibly have exposed the pharynx to N gonorrhoeae was examined (table 1).

Results

From June 2001 to December 2004, a total of 1427 participants were enrolled. The median age at enrolment was 35 years (range 18–75 years), 68% were Australian-born, and 74% were of Anglo ethnicity. Ninety-five per cent of participants self-identified as gay or homosexual.

By the end of the HIM study in June 2007, the total follow-up time was 4537.8 PY, and median follow-up time for participants was 3.9 years. The follow-up time from commencement of NAAT testing in January 2003 to the end of the HIM study in June 2007 was 3781.5 PY. A total of 4090 NAAT N gonorrhoeae pharyngeal tests (90.2% of eligible visits) were performed in this time period. Younger participants were more likely to decline testing (mean age 37.5 vs 39.9 years, p<0.001). Those declining testing reported less frequent oral sexual behaviours. In addition to having fewer male partners in the last 6 months (p<0.001), they reported less frequent insertive oro-anal sex with both regular (p=0.015) and casual (p=0.005) partners, and less frequent receptive penile-oral sex without ejaculation with regular partners (p=0.041). Those who declined testing were also non-significantly less likely to report receptive penile-oral sex with ejaculation with casual partners (p=0.057) and sexual contact with gonorrhoea in the past year (p=0.087).

Diagnosis

Sixty-five of 223 pharyngeal samples initially reactive on the BD ProbeTec assay tested positive for N gonorrhoeae with the LC porA assay.

Prevalence

At baseline, 2.41% (95% CI 1.68 to 3.35%) of the men reported a diagnosis of pharyngeal gonorrhoea in the last 12 months. On initial testing (n=1225), seven pharyngeal N gonorrhoeae infections were identified, a prevalence of 0.57% (95% CI 0.23 to 1.17%).

Incidence

Study-visit diagnoses

Sixty-five pharyngeal N gonorrhoeae infections were diagnosed at study visits; an incidence rate of 1.51 per 100 PY (95% CI 1.19 to 1.93).

Interval diagnoses

During the entire HIM study period, there were 161 self-reported pharyngeal N gonorrhoeae infections in the previous 12 months; an incidence rate of 2.75 per 100 PY (95% CI 2.36 to 3.21).

Combined diagnoses

The combined estimate could be calculated for interviews completed from January 2003 onwards. No participants had both self-reported and NAAT confirmed pharyngeal gonorrhoea diagnoses in a single year. There were 193 combined study visit or self-reported pharyngeal N gonorrhoeae infections; an incidence rate of 4.45 per 100 PY (95% CI 3.86 to 5.12). For the combined incidence estimate, the majority (128/193, 66.3%) of pharyngeal N gonorrhoeae infections were interval diagnoses.

Of sixty-five infections identified at study visits, 55 (84.6%) participants had N gonorrhoeae only in the pharynx, nine (13.8%) had concurrent anal infection, and one (1.5%) had concurrent urethral N gonorrhoeae infection. This contrasts with anogenital gonorrhoea where over one-third of participants had concurrent gonorrhoea infections diagnosed at other sites. Among nine participants with urethral gonorrhoea, one pharyngeal (11.1%) and two rectal (22.2%) infections were identified. Among 32 participants with rectal gonorrhoea, two urethral (6.3%) and nine pharyngeal (28.1%) infections were identified. Only two participants were diagnosed as having pharyngeal gonorrhoea on two occasions at annual HIM study testing. For both these participants, infections were diagnosed 2 years apart, and both had a negative pharyngeal N gonorrhoeae test at the annual visit between positive tests.

There was no association of pharyngeal gonorrhoea with pharyngeal symptoms in the previous week. Sore throat was reported at 18.1% of visits in those without infection and 15.4% of visits in those with infection (OR 0.82, 95% CI 0.39 to 1.73, p=0.609).

On univariate analysis, pharyngeal N gonorrhoeae was significantly associated with younger age and sexual contact with somebody known to have gonorrhoea (table 2).

Table 2

Univariate analysis of demographic risk factors and sexual contact with gonorrhoea for incident pharyngeal N gonorrhoeae infection in the HIV-negative HIM cohort

Regular partners: univariate analysis (data not shown)

Among men reporting regular partners, infection was significantly associated with an increasing number of male partners in the previous 6 months (p<0.001), having sexual partners outside the relationship (p<0.001) and group sex involving a regular partner (p=0.001). Of sexual behaviours with regular partners, only insertive oro-anal sex (‘rimming’) was associated with pharyngeal N gonorrhoeae (p=0.036). Apart from this single risk factor, risk was therefore related to sexual behaviours with casual partners.

Casual partners: univariate analysis

In relation to sexual contact with casual partners, those who reported a higher number of male sexual partners in the last 6 months were at increased risk of infection. Sexual behaviours with casual partners that were significantly associated with pharyngeal N gonorrhoeae infection on univariate analysis were wet and dry kissing, receptive penile-oral sex with and without ejaculation and insertive oro-anal sex (table 1).

Regular partners: multivariate analysis (data not shown)

In both multivariate models, incident pharyngeal N gonorrhoeae was significantly associated with younger age, sexual contact with somebody known to have gonorrhoea and reporting more male partners in the last 6 months. The only sexual behaviour with regular partners that approached statistical significance was the association of more frequent insertive oro-anal sex with incident pharyngeal N gonorrhoeae infection (p trend=0.089).

Casual partners: multivariate analysis

Reporting more frequent insertive oro-anal sex with casual partners was significantly associated with incident pharyngeal N gonorrhoeae infection (p trend=0.044). Although not reaching statistical significance, compared with participants ‘never’ practising receptive penile-oral sex without ejaculation with casual partners, adjusted HRs were over 3 for those reporting this practice ‘occasionally’ and ‘often’ (table 3).

Table 3

Multivariate analysis of demographic factors, gonorrhoea contact and sexual behavioural risk factors with casual partners in last 6 months for pharyngeal N gonorrhoeae infection in the HIV-negative HIM cohort among men who reported casual partner(s)

Discussion

The baseline prevalence of study-visit-diagnosed pharyngeal N gonorrhoeae infection in this community-based cohort of HIV-negative homosexual men was 0.57%, and the incidence of infection detected on study-visit testing was 1.51 per 100 PY. The combined incidence of pharyngeal gonorrhoea, including self-reported infections, was 4.45 per 100 PY. The acquisition of pharyngeal gonorrhoea was strongly associated with sexual contact with someone known to have gonorrhoea, younger age and increasing numbers of male sexual partners in the past 6 months. Participants who reported more frequent insertive oro-anal sex (‘rimming’) with casual partners were at increased risk of infection, while the association with more frequent insertive oro-anal sex with regular partners was of borderline significance.

There are few incidence data on pharyngeal gonorrhoea infections among homosexual men. The one other prospective community-based study reported a higher incidence, of 11.2 per 100 PY among HIV-negative MSM.4 The BD ProbeTec assay was also used in that study, but no supplemental gonococcal NAAT testing was performed. HIM study data have suggested, at least in an Australian homosexual population, that the BD ProbeTec assay suffers from a low positive predictive value (PPV) for the diagnosis of pharyngeal gonorrhoea. The PPV of the BD ProbeTec assay for pharyngeal N gonorrhoeae in HIM was only 29.1% (95% CI 23.3% to 35.6%). Poor performance of the ProbeTec assay for pharyngeal gonorrhoea diagnosis has also been described among MSM in a clinic-based setting.21 Thus, relying solely on the results of the BD ProbeTec assay could lead to a substantial overestimation of gonorrhoea incidence.

The combined study visit and interval incidence of urethral and anal gonorrhoea in HIM was 3.49 and 2.96 per 100 PY, respectively,14 lower than the combined incidence of pharyngeal gonorrhoea. Higher rates of pharyngeal relative to anogenital gonococcal infections among MSM have recently been reported in community-based settings by others.3 4 This may partly reflect the lack of association of symptoms with pharyngeal gonorrhoea compared with anogenital infections which more commonly lead to clinical presentation.21 This lack of association of sore throat with infection among HIM participants supports findings from other community-based3 4 and clinic-based9 21 22 MSM populations. The relatively lower prevalence of unprotected anal as compared with unprotected oral sex that has been described in homosexual men13 is likely to be another reason why pharyngeal infection is more common than infection at other sites.

The association of pharyngeal gonorrhoea with insertive oro-anal sex reported here has not previously been described or investigated prospectively. The only study to assess this behaviour was performed on a cross-sectional sample of MSM clinic attendees21 and found no association. The fact that pharyngeal gonorrhoea was associated with more frequent insertive oro-anal sex with casual (p trend=0.044) as well as regular (p trend=0.089) partners in our study strengthens the likelihood that this is a real finding. This has biological plausibility, as the gonococcus is detected more frequently in the anus than the urethra of MSM in community-based samples.3 4 In HIM, the baseline prevalence of anal gonorrhoea was almost three times that of urethral gonorrhoea, while the incidence of study visit anal gonorrhoea was almost fourfold greater than urethral gonorrhoea during the study period.14 Undoubtedly, the stronger association of symptoms with urethral compared with anal gonorrhoea21 contributes to a longer duration and thus a higher prevalence of anal gonorrhoea infection among MSM. Should the gonococcus be even moderately efficiently transmitted via oro-anal sex, this could plausibly explain the stronger association of pharyngeal gonorrhoea with oro-anal compared with penile-oral sex.

Several studies1 4 9 21 have reported an independent association of receptive penile-oral sex with pharyngeal gonorrhoea infection among MSM. Although the relative risks associated with this behaviour in the HIM cohort were in the order of 3, it was not statistically significant. The wide CIs for penile-oral sex without ejaculation were because the referent category (reporting the behaviour ‘never’ in each 6-month period) was rare (table 3) leading to limited power. Our lack of observed association should not, we believe, be taken to contradict the evidence from these other observational studies that receptive penile-oral sex is a major route of transmission of gonorrhoea to the pharynx.

This study had at least three important strengths. First, this community-based cohort is likely to be broadly representative of gay community-attached urban HIV-negative men in Australia. Second, unlike previous studies4 5 the strict diagnostic criteria for pharyngeal N gonorrhoeae involved supplementary testing of all positive BD ProbeTec pharyngeal gonorrhoea samples with a highly sensitive and specific assay targeting the gonococcal porA pseudogene.18 Third, the study combined both self-reported diagnoses of pharyngeal gonorrhoea in the last year and NAAT-confirmed diagnoses at the time of the annual visits. Relying solely on NAAT-confirmed study-visit diagnoses would substantially underestimate the incidence of pharyngeal gonorrhoea in HIM.

A possible weakness of these data is that the estimate of interval diagnoses was dependent upon participants' self-report, and validation against their medical record was not attempted. In addition, the Australian diagnostic standard for pharyngeal gonorrhoea testing in MSM is culture, which has a lower sensitivity than pharyngeal NAAT diagnosis of gonorrhoea.22 This may have resulted in an underestimate of self-reported pharyngeal gonorrhoea incidence during the study period. However, it was reassuring that risk factors for study-visit diagnoses were very similar to those for interval diagnoses (data not shown). It is possible that the sensitivity of the BD ProbeTec assay was higher than that of the porA assay for some samples, leading to misclassification of some test results as false positive. However, because of the mobility of the gay population in Australia, it is unlikely that the excellent porA assay sensitivity obtained in another eastern Australian city for non-genital samples would be substantially lower among Health in Men Study participants tested in Sydney.17 The fact that that the incidence of pharyngeal gonorrhoea was almost threefold higher than the prevalence suggests that the average duration of infection was short. It is likely that many pharyngeal gonorrhoea infections spontaneously resolved in study participants and therefore went undetected. Although no natural history studies have been performed using NAAT diagnosis, spontaneous resolution of pharyngeal gonorrhoea appears to be common,23 with most cases culture-negative after a week and all untreated cases resolving within 3 months.24 An additional explanation is that participants' antibiotic use for other infections may have led to resolution of pharyngeal gonorrhoea infections. However, data on antibiotic use were not collected as part of the community-based HIM study. A further possible limitation is that only sexual behaviours directly involving the oropharynx were analysed as risk factors for pharyngeal gonorrhoea infection. Behaviours such as digital/oral contact were not assessed in the study.

Three novel findings of our analysis of pharyngeal gonorrhoea in this cohort of Sydney men should inform pharyngeal screening guidelines in MSM: the large proportion of pharyngeal gonorrhoea (almost 85%) which would have been missed by screening only anogenital sites in this community-based study; the independent association of pharyngeal gonorrhoea with insertive oro-anal sex; and the likelihood that more frequent screening of the pharynx for gonorrhoea would detect substantially more infections. However, it is possible that spontaneous resolution of NAAT-detected pharyngeal gonococcal infection occurs at a similar rate to that of culture-detected pharyngeal gonorrhoea.23 24 Further studies are needed to establish the natural history and sequelae of NAAT-detected pharyngeal gonorrhoea before an increase in frequency of pharyngeal gonococcal screening in MSM can be recommended. Nonetheless, the incidence of infection was high in the HIM cohort, despite relatively infrequent screening. In addition, most cases were identified as isolated infections in the pharynx, underscoring the importance of regular pharyngeal gonorrhoea screening in MSM as recommended in Centers for Disease Control (CDC) guidelines.25 However, these guidelines only advocate screening MSM who engage in ‘receptive oral intercourse,’ and the association of pharyngeal gonorrhoea with insertive oro-anal sex in our cohort suggests that a broader range of oral sexual behaviours are involved in gonococcal transmission to the pharynx. The pharynx as a source of anogenital gonorrhoea has been suggested by several studies15 26 27 including HIM14 which have described a range of oral sexual practices to be predictive of gonorrhoea at anogenital sites. Thus, routine screening and treatment of pharyngeal gonorrhoea is an important component of gonorrhoea control programmes in MSM. In combination with other public health measures including repeat testing at 3 months and partner notification, pharyngeal gonorrhoea screening is likely to be important in reducing anogenital gonorrhoea prevalence among MSM.

Key messages

  • The high incidence-to-prevalence ratio of pharyngeal N gonorrhoeae in this community-based cohort of Australian homosexual men suggests frequent spontaneous resolution of NAAT-detected infection.

  • Almost 85% of pharyngeal gonorrhoea infections were identified in the absence of concurrent anogenital infection.

  • This is the first study to describe an independent association of pharyngeal gonorrhoea with rimming in MSM.

Acknowledgments

We thank all the participants, the dedicated HIM Study team and the participating doctors and clinics. We also thank Becton Dickinson for providing testing materials for gonorrhoea and chlamydia.

References

View Abstract

Footnotes

  • Funding During the study, DJT was supported by National Health and Medical Research Council Public Health Scholarship no 351044 and, subsequently, the Royal Australasian College of Physicians GSK Scholarship for virological research. The National Centre in HIV Epidemiology and Clinical Research and the National Centre in HIV Social Research are funded by the Australian government Department of Health and Ageing. The Health in Men Cohort study was funded by the National Institutes of Health, a component of the US Department of Health and Human Services (NIH/NIAID/DAIDS: HVDDT Award N01-AI-05395), the Australian government Department of Health and Ageing (Canberra) and the New South Wales Health Department (Sydney), and the National Health and Medical Research Council (project grant no 400944).

  • Competing interests None.

  • Ethics approval Ethics approval was provided by the University of New South Wales.

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

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.