Contribution of transmission in HIV-positive men who have sex with men to evolving epidemics of sexually transmitted infections in England : an analysis using multiple data sources , 2009 – 2013

Citation style for this article: Malek R, Mitchell H, Furegato M, Simms I, Mohammed H, Nardone A, Hughes G. Contribution of transmission in HIV-positive men who have sex with men to evolving epidemics of sexually transmitted infections in England: an analysis using multiple data sources, 2009–2013. Euro Surveill. 2015;20(15):pii=21093. Available online: http://www.eurosurveillance.org/ViewArticle.aspx?ArticleId=21093


Surveillance and outbreak reports
Contribution of transmission in HIV-positive men who have sex with men to evolving epidemics of sexually transmitted infections in England: an analysis using multiple data sources, 2009-2013

Introduction
The United Kingdom (UK) has seen a steady increase in diagnoses of sexually transmitted infections (STIs) in the last decade, in particular, in men who have sex with men (MSM). From 2012 to 2013, gonorrhoea and syphilis diagnoses in MSM increased by 26% and 12% respectively [1]. The emergence of resistance and reduced sensitivity to frontline treatments of gonorrhoea is of global concern and may have contributed to high levels of gonorrhoea transmission in MSM [2][3][4]. Additionally, outbreaks of less common STIs such as Lymphogranuloma venereum (LGV) [5], together with shigellosis (which can be sexually transmissible) [6] in this population are of particular concern.
There is increasing evidence that HIV-positive MSM in the UK are disproportionately affected by STIs. Recent data demonstrated that of MSM diagnosed with syphilis, 35% were HIV positive [7]. HIV-positive MSM have also been found to account for more than 80% of LGV cases [5]. In part, these observations may reflect seroadaptive behaviours in HIV-positive MSM [8], in which different sexual practices such as unprotected (i.e. condomless) anal intercourse are adopted according to the reported HIV status of both partners in order to reduce the risk of transmitting or acquiring HIV. Seroadaptive behaviours vary considerably and include serosorting (limiting sexual partners to those with the same HIV status as themselves), strategic positioning (adopting a specific sexual position according to the HIV status of one's partner), withdrawal before ejaculation, and negotiating around viral load [9]. However, seroadaptive behaviours may come at a cost of increased transmission of STIs [10][11][12].
The UK has a growing population of HIV-positive MSM [13] and recent studies suggest that they may be increasingly adopting seroadaptive behaviours [8,14]. In this study we used multiple surveillance data sources and population survey data to estimate the risk of STIs in HIV-positive MSM and assessed whether transmission in HIV-positive MSM has contributed to recent STI epidemics in England.

Sources of data
Descriptive and comparative data analyses of the incidence of STIs were undertaken using data from GUMCADv2 (Genitourinary Medicine Clinic Activity Dataset) [15]. The Survey of Prevalent HIV Infections Diagnosed (SOPHID) [16], the third National Survey of Sexual Attitudes and Lifestyles (Natsal-3) [17] and census data from the Office for National Statistics (ONS) [18] were employed to estimate denominator populations.
GUMCADv2 is a mandatory electronic pseudoanonymised (i.e. contains the sex, age and hospital/ clinic number of each patient but no patient identifiable information such as name, date of birth or postcode of residence is included) [19] patient-level dataset submitted to Public Health England (PHE) by all genitourinary medicine (GUM) clinics in England. The dataset contains information on all STI diagnoses and services provided for each patient as well as information on patient demographic such as sexual orientation, age, sex, ethnicity, area of residence and country of birth [19]. A unique patient identifier is assigned to each patient attending a given GUM clinic, allowing subsequent visits by the same patient to the same clinic to be identified. SOPHID is a cross-sectional survey of all persons with diagnosed HIV infection who attend for HIV care at an NHS site in England, Wales and Northern Ireland. Age, sex, probable route of HIV infection, ethnicity, antiretroviral therapy (ART) status, CD4 cell count, region of residence and region of care provider are measured for each calendar year.
Natsal-3 is a nationally representative survey conducted between 2010 and 2012 on 15,162 individuals and provides information on key sexual behaviours, risk factors and also includes biological sampling and testing.
ONS is the national statistical institute for the UK and is responsible for collecting and publishing statistics related to the economy, population and society at national, regional and local levels.

Study population
MSM were defined as men who reported a homosexual or bisexual orientation at least once over the study period. MSM were defined as diagnosed HIV positive (hereafter referred to as 'HIV-positive' MSM) if they were diagnosed with HIV at least six weeks before their STI diagnosis, as newly diagnosed with HIV if they were diagnosed within six weeks of their STI diagnosis, and as HIV-negative/undiagnosed if there was no evidence of an HIV diagnosis in their GUMCADv2 record. For the analysis of acute STIs, STI population rates and association between STI outcomes and HIV status, MSM with new HIV diagnoses were grouped with those of negative or unknown status and referred to collectively as 'HIV-negative/undiagnosed', as the newly diagnosed men were assumed to be undiagnosed at the time of their STI exposure.

Data analysis Acute sexually transmitted infections in known HIV-positive men who have sex with men
Episodes of acute STI diagnoses including gonorrhoea (acute and complicated), syphilis (primary, secondary and early latent), chlamydia, genital warts (first episode) and genital herpes (first episode) and HIV status in MSM were identified using Sexual Health and HIV Activity Property Type (SHHAPT) codes from GUMCADv2 for the years 2008 to 2013. Acute gonorrhoea includes all new cases of uncomplicated gonorrhoea of the lower genitourinary tract, anorectum, mouth, throat and adult conjunctivitis; complicated gonorrhoea includes all upper genitourinary tract complications (such as pelvic inflammatory disease and epididymitis) and systemic complications [20]. The proportion of STI diagnoses which were in HIV-positive MSM was calculated for each STI by year.

Sexually transmitted infection population rates in HIV-negative/undiagnosed and HIV-positive men who have sex with men
The rates of acute bacterial (gonorrhoea, chlamydia and syphilis) and acute viral STIs (first episode of genital warts and first episode of genital herpes) in HIV-positive and HIV negative/undiagnosed MSM were compared from 2009 to 2013. The numerators were individual HIV-positive and negative/undiagnosed MSM presenting at GUM clinics each year using data from GUMCADv2. The denominator for HIV-positive MSM was identified from SOPHID. The proportion of men aged 15 to 74 years who are MSM was estimated using data from Natsal-3. An estimated 2.6% of men in the UK had at least one male sexual partner in the past five years across all age groups of men [21]. This was applied to the mid-year population estimates from the ONS for the number of men aged 15 to 74 years for each year [22,23], and the estimated number of HIV-positive MSM was subtracted from this to calculate the denominator for HIV-negative/undiagnosed MSM. The results of a sensitivity analysis (data not shown) showed a small effect when using the upper and lower limits of the confidence interval (2.1% and 3% respectively) of the estimation of men in the UK having at least one male sexual partner in the past five years across all ages of men. For the most recent year of analysis, the range for bacterial STIs in HIV-negative MSM was 33 per 1,000 to 48 per 1,000.

Associations between sexually transmitted infection outcomes and HIV status
Univariate and multivariate logistic regression models were prepared for the MSM population attending GUM clinics in England for each individual MSM diagnosed with an acute bacterial STI (gonorrhoea, chlamydia and syphilis) for each year from 2009 to 2013 inclusive. Explanatory variables included all demographic variables from GUMCADv2 (age, ethnicity, continent of birth, and area of residence), diagnosed HIV status, and number of tests in the previous 12-month period for each respective bacterial STI; all were included in the final multivariate model.

Gonorrhoea and chlamydia re-infection rates by HIV status
The probability of patients who became re-infected with the most common bacterial STIs (gonorrhoea and chlamydia) within one year was estimated by the Kaplan-Meier method. Patients became at-risk from 42 days after the time of first attendance with gonorrhoea and chlamydia [24] and were censored at the end of the study period (31 December 2013). Data for calculating re-infection rates was obtained from GUMCADv2.
All statistical analyses were undertaken using Stata version 12 (StataCorp, College Station, Texas, US). P values < 0.05 were considered to be statistically significant.

Results
During the study period, the number of MSM attending GUM clinics recorded in GUMCADv2  Figure 1. Overall, from 2009 to 2013, this proportion increased for all acute STIs: from 25% to 40% for syphilis, 16% to 25% for chlamydia, 15% to 24% for gonorrhoea, 19% to 21% for genital herpes and 7% to 10% for genital warts.

Sexually transmitted infection population rates in HIV-negative/undiagnosed and HIV-positive men who have sex with men
The rate of acute bacterial STIs in HIV-positive MSM nearly trebled (64 per 1,000 to 161 per 1,000), and increased from 3.2 times higher than the rate in HIVnegative/undiagnosed MSM in 2009 to 4.2 times higher in 2013 ( Figure 2). For HIV-negative/undiagnosed MSM, the rate of acute bacterial STIs also increased to a lesser degree (19 per 1,000 to 38 per 1,000). The rate of acute viral STIs in HIV-positive MSM was approximately twice that observed in HIV-negative/undiagnosed MSM and remained fairly stable over the five year period (15 to 18 per 1,000 in HIV-positive and 8 to 9 per 1,000 in negative/undiagnosed MSM). This indicates that the rate of acute STIs in HIV-positive MSM was close to one in five (179 per 1,000).  Table 1 shows the final multivariate logistic regression models for gonorrhoea, chlamydia and syphilis adjusted for all explanatory variables. Data are presented for the most recent available year only. Table 2 shows the change in the adjusted odds ratio over time by HIV status. When compared with HIV-negative/undiagnosed MSM, the adjusted odds ratio (aOR) of being diagnosed with gonorrhoea, chlamydia and syphilis was significantly higher in HIV-positive MSM in all years analysed and increased over time (p<0.001).

Gonorrhoea and chlamydia re-infection rates by HIV status
The estimated probability of gonorrhoea and chlamydia re-infection is shown in Figure 3.    [25], and similar studies in Europe [26,27]. A study of LGV re-infection in the UK also found that, at baseline, repeaters were more likely to be HIVpositive compared with non-repeaters [28].
It seems probable that these changes in STI transmission patterns in HIV-positive MSM reflect increasing adoption of HIV seroadaptive behaviours and their impact on sexual network structures. HIV-positive MSM reporting seroadaptive behaviours are at higher risk of STIs compared with HIV-negative MSM [10][11][12].
One study reported that HIV-positive MSM practising seroadaptive behaviours had a threefold increased risk of bacterial STIs, with almost a third of HIV-positive MSM reporting an STI in the past year, compared with 9% of HIV-negative MSM [11]. However, despite the considerable changes in STI transmission patterns in HIVpositive MSM seen in our study, the degree to which transmission in HIV-positive MSM engaging in seroadaptive behaviours is fuelling current STI epidemics in England is unclear. The recent emergence of relatively rare infections such as LGV and S. flexneri in the UK has been strongly and predominantly associated with transmission in HIV-positive MSM [5,6,29]. Likewise for syphilis, our study showed that 39.1% of cases in MSM in England were known to be HIV-positive, and this is consistent with data from the United States, Australia and Europe [26,27,30,31]. The duration of infectiousness with Shigella, LGV and syphilis may be short as these infections are typically symptomatic, so it is highly likely that their transmission is being sustained in highly active sexual networks of HIV-positive MSM engaging in seroadaptive behaviours [32]. However, fewer than 25% of MSM diagnosed with chlamydia and gonorrhoea (and even less with viral STIs) in our study were HIV-positive. Therefore, while seroadaptive behaviours in HIV-positive MSM may be making an important contribution to the transmission of chlamydia, gonorrhoea and viral STIs in MSM, they are not necessary to sustain infection at endemic levels in the wider MSM population.
HIV seroadaptive behaviours will likely have other negative or unintended consequences for the sexual health of MSM. The presence of an STI may compromise the health of HIV-positive MSM through several mechanisms including a reduction in CD4 cell count as well as acute increases in HIV viral load, which may compromise effective antiretroviral therapy [33][34][35].
In addition, STIs may also increase HIV infectiousness by facilitating HIV shedding in the genital tract or rectal mucosa [36]. Further, evidence suggests that a low viral load may reduce the probability of infecting Table 2 Adjusted  [8,14].
A study by Fox et al. showed significant reduction in self-reported HIV transmission-risk behaviour in MSM recently diagnosed with HIV, with patients reporting greater condom use and fewer sexual partners [40].
However, these may have limited impact on STI incidence due to the various routes of transmission of STIs (such as oral, digital and use of sex toys). Furthermore, in MSM recently diagnosed with HIV, those reporting continued transmission-risk behaviour were more likely to have another STI [40]. Thus, health promotion activities should also consider the broader context around sexual risk-taking in MSM, especially in those diagnosed with HIV. The recent Public Health England framework for promoting the health and well being of MSM highlighted the interaction of mental health, alcohol and drug use, and sexual risk behaviour [41]. There is increasing concern on the interaction between drug use and STIs, especially in HIV-positive  MSM, and that in developing appropriate interventions and services, the specific needs of HIV-positive MSM should be considered [29]. Furthermore, the high incidence of asymptomatic STIs, especially in extra-genital sites [42,43] emphasises the need to promote regular screening for STIs in HIV-positive individuals [44,45]. Nonetheless, surveys of healthcare providers in the US have reported significant barriers, especially in screening for gonorrhoea and chlamydia, which include time constraints, difficulty obtaining a sexual history, language and cultural barriers, and patient confidentiality concerns [46].
There are several limitations of this study. There may be a degree of ascertainment bias in the assessment of HIV status in MSM in patients with a longer history in GUMCADv2. However, sensitivity analysis (data not shown) was performed by identifying HIV-positive MSM from a retrospective review of a single year of GUMCADv2 data, and there was only a minimal impact on the results. The probability of repeat infection may also have been underestimated as, in this analysis, repeat diagnosis was used as a proxy measure for repeat infection. Some patients will have become re-infected but will remain undiagnosed, as only patients who returned to the same clinic for testing were assessed in the analysis. This is a limitation of GUMCADv2, as it allows only longitudinal patient data within a particular clinic or service and attendances by the same patient at different clinics cannot be monitored [19]. However, the data quality and completeness of GUMCADv2 is extremely high with 100% submissions from GUM clinics [19]. HIV-positive MSM engaged in care are also more likely to return regularly for STI screening during clinic visits and this may have contributed to the proportion and rates of STIs observed. Diagnosis of STIs (including HIV) is dependent on screening practices and frequency of screening. It is therefore not a true measure of incidence of infection but provides a good proxy for infection. A further limitation is the estimation of the true size of the MSM population in the UK. However, the methodology employed in Natsal-3 is among the most robust to estimate the size of this population. Furthermore, between Natsal-2 and Natsal-3, there was no significant increase in the proportion of men reporting same sex partners in the past 5 years [21]. Thus, the results of this longitudinal data analysis provide valuable insights into the complexity and evolution of STI epidemics in England.
The presence of an increasing proportion and rates of acute STIs in HIV-positive MSM, a population which also has higher rates of repeat infection and reports higher risk sexual behaviour, presents an increased risk for the sexual health of all MSM. This therefore merits public health action through improved monitoring and intervention. Currently, the collection of behavioural data is being piloted alongside that of clinical and socio-demographic data and this will allow further insights into the impact of seroadaptive practices to be explored in the future. Improved public health interventions with a holistic approach focussing on promoting condom use, reducing high risk behaviour and increasing the frequency of STI testing in MSM should be a priority.