Recreational drug use, polydrug use, and sexual behaviour in HIV-diagnosed men who have sex with men in the UK: results from the cross-sectional ASTRA study

of the study We describe patterns of drug use and associations with sexual behaviours in HIV-diagnosed in the UK. We used data from the cross-sectional ASTRA study, which recruited participants aged 18 years or older with HIV from eight HIV outpatient clinics in the UK between Feb 1, 2011, and Dec 31, 2012. We examined data for MSM, assessing the prevalence of recreational drug use and polydrug use in the previous 3 months and associations with sociodemographic and HIV-related factors. We examined the association of polydrug use with measures of condomless sex in the previous 3 months and with other sexual behaviours. analysis data for 2248 MSM: 2136 (95%) were gay, 1973 (89%) were white, 1904 (85%) were on antiretroviral treatment (ART), and 1682 (76%) had a viral load of 50 copies per mL or lower. 1138 (51%) in the previous 3 months; 608 (27%) nitrites, 477 (21%) used 460 (21%) 453 280 (13%) ketamine, 258 (12%) 3,4-methylenedioxy-N-methylamphetamine (MDMA), 221 (10%) used gamma-hydroxybutyrate or gamma-butyrolactone, 175 (8%) used methamphetamine, and 162 (7%) used mephedrone. In the 1138 individuals who used drugs, 529 (47%) used three or more drugs and 241 (21%) used ﬁ ve or more. Prevalence of injection drug use was 3% (n=68). Drug use was independently associated with younger age (p<0·0001), not being religious (p=0·001), having an HIV-positive stable partner (p=0·0008), HIV-serostatus disclosure (p=0·009), smoking (p<0·0001), evidence of harmful alcohol drinking (p=0·0001), and ART non-adherence (p<0·0001). Increasing polydrug use was associated with increasing prevalence of condomless sex (prevalence range from no drug use to use of ﬁ ve or more drugs was 24% to 78%), condomless sex with HIV-seroconcordant partners (17% to 69%), condomless sex with HIV-serodiscordant partners (10% to 25%), and higher-HIV-risk condomless sex after taking viral load into account (4% to 16%; p≤0·005 for all). Associations were similar after adjustment for sociodemographic and HIV-related factors. Methamphetamine was more strongly associated with higher-HIV-risk condomless sex than were other commonly used drugs. Interpretation Polydrug use is prevalent in HIV-diagnosed MSM and is strongly associated with condomless sex. Specialist support services for MSM with HIV who use recreational drugs might be beneﬁ cial in the reduction of harm and prevention of ongoing transmission of HIV and other sexually transmitted infections. partners accounting for viral load, which has not previously been examined. Our ﬁ ndings show that recreational drug use and polydrug use are prevalent, and strongly associated with condomless sex and multiple new sexual partners among HIV-diagnosed MSM. These ﬁ ndings draw attention to the importance of tailored harm reduction support for MSM who use recreational drugs and to the need to address the prevention of HIV and sexually transmitted infection in this group.


Introduction
Recreational drug use is an important public health concern in men who have sex with men (MSM) in the UK. 1,2 Findings from studies done in Europe and the USA show that drug use is more prevalent in MSM compared with the general population. [3][4][5] Drugs that are usually taken in nightclubs and in unlicensed dance parties are collectively known as club drugs and encompass controlled and non-controlled substances including 3,4-methylenedioxy-N-methylamphetamine (MDMA, or ecstasy), methamphetamine (crystal meth), gamma-hydroxybutyrate (GHB), gamma-butyrolactone (GBL), mephedrone, and ketamine. 6,7 Although fi ndings from many studies in the past 15 years have shown the high prevalence of drug use in MSM, [4][5][6] club drug use might have increased in this group possibly becoming normalised within sexual contexts. 1,8 One London club drug clinic (a free health service for adult clubbers and lesbian, gay, bisexual, and transgender people), for which data have been collected since 2005, reports increasing use of methamphetamine, mephedrone, and GHB, solely in facilitating sex (known as chemsex) in MSM attending the clinic, 2 and it has been suggested that the practice of injecting recreational drugs at sex parties might be increasing in MSM. 9,10 A possible changing culture of recreational drug use in MSM in the UK might be linked to increased sexual risk behaviours. 1 This increase could in turn lead to an increase in new sexually transmitted infections (STIs) including HIV, the transmission of which is high in MSM in the UK (3250 new diagnoses in 2012). 11 Findings from many online and gay-venue-based studies in MSM in Europe, Australia, and North America have associated use of some drugs (such as methamphetamine and erectile dysfunction drugs) with sex without a condom (herein referred to as condomless sex), including with HIV-serodiscordant status partners. 5,[12][13][14][15][16] Findings from several studies have shown HIV-diagnosed MSM to be more likely to use almost all types of recreational drugs compared with MSM who are HIV-negative or un diagnosed. 4,14,17,18 Polydrug use (use of more than one drug at the same time or within the same time period) might indicate a more severe substance abuse problem. Findings from some studies from the UK and USA suggest that polydrug use is particularly prevalent in HIV-positive MSM. 4,19 Additionally, polydrug use was linked to condomless sex with casual partners in a US study of HIV-diagnosed MSM surveyed in bars and sex clubs. 20 However, fi ndings relating to HIV-diagnosed MSM from online and venue-based studies might have restricted generalisability. UK researchers 21 reported the prevalence of methamphetamine and other recreational drug use in a sample of HIV-diagnosed MSM from a London HIV clinic between 2002 and 2003. 21 Since then, few studies have assessed patterns of recreational drug use and the extent of polydrug use from representative samples of HIV-diagnosed MSM in the UK.
We describe the prevalence of recreational drug use and polydrug use, their association to socio-demographic, lifestyle, and HIV-related factors, and their relation with condomless sex and other sexual behaviours in HIVdiagnosed MSM in the UK from the ASTRA (Antiretrovirals, Sexual Transmission Risk and Attitudes) study.

Study design
The ASTRA study has been described elsewhere. 22 Briefl y, ASTRA recruited men and women aged 18 years or older with HIV who were attending eight HIV outpatient clinics in the UK between Feb 1, 2011, and Dec 31, 2012. 22 Participants completed a confi dential, self-administered questionnaire that sought information about sociodemographic factors ( 23 and recreational drug use), and sexual behaviour measures. Consent to participate included permission to collect latest CD4 count and HIV plasma viral load. Men who identifi ed as being gay or bisexual, or who reported sex with men in the previous 3 months, were classifi ed as MSM. The study was approved by the North West London REC 2 research ethics committee (ref 10/H0720/70).

Procedures
Participants were asked if they had used recreational drugs in the past 3 months and if so which ones. Recreational drugs included the following: acid, lysergic acid diethylamide (LSD), or magic mushrooms (all grouped as psychedelics); anabolic steroids; cannabis (marijuana); cocaine (coke); crack; codeine; crystal meth (methamphetamine); ecstasy (E); GHB (liquid ecstasy); heroin, ketamine (K); khat (chat); mephedrone; morphine; opium; poppers (amyl nitrites); speed (amphetamine); and erectile dysfunction drugs sildenafi l and tadalafi l. Other drugs were coded to the above categories according to slang names. 24 Polydrug use was assessed by the number of diff erent drugs used during the previous 3 months. Club drugs were defi ned as MDMA, GHB (liquid ecstasy) or GBL, ketamine, and mephedrone. We did not include methamphetamine in the defi nition for club drugs because reports suggest it might be more commonly used in private sex parties. 1,10,25 Participants were asked about injecting recreational drugs and needle sharing in the past 3 months.
We defi ned ten sexual behaviour measures in the previous 3 months (unless otherwise stated): any anal or vaginal sex, condomless sex (anal or vaginal sex without a condom), condomless sex with a seroconcordant (HIVdiagnosed) partner, condomless sex with a serodiscordant partner (of unknown or HIV-negative status), higher-HIVrisk condomless sex with a serodiscordant partner (if the participant was either not on ART, had latest HIV viral load greater than 50 copies per mL, or had a new STI diagnosis in the previous 3 months), new STI diagnosis (syphilis, gonorrhoea, chlamydia, lymphogranuloma venereum, new hepatitis B and C, genital herpes, genital warts, trichomonas, non-specifi c urethritis, non-gonoccocal urethritis), group sex (with more than one person on the same occasion), used the internet to fi nd sexual partners, agreement to the statement "I am less likely to use a condom with a casual partner (no recall period)", and number of new sexual partners in the previous year.

Statistical analysis
Our analysis includes HIV-diagnosed MSM only. First, we assessed prevalence of any recreational drug use, and use of one, two, three, four, or fi ve or more types of drug in the past 3 months, and examined patterns of drug use according to number of drugs used. Second, we examined the association of sociodemographic, HIV-related, and lifestyle factors with any recreational drug use (dependent variable); in recreational drug users, we assessed factors associated with use of four or more drugs versus one to three drugs. We examined crude associations with χ² tests, and used modifi ed Poisson regression models with robust error variances 26 to produce prevalence ratios (PRs) with 95% CIs. Factors with p<0·15 in univariable analysis were considered in multivariable models; factors with p<0·05 were retained in the fi nal model. Third, we examined associations of recreational drug use and polydrug use (none, one, two, three, four, or fi ve or more drugs) with each of the ten measures of sexual behaviour (dependent variables). We assessed unadjusted associations with χ² tests and adjusted (for signifi cant sociodemographic and HIV-related factors) associations with modifi ed Poisson regression.
Because some recreational drugs are reported to be used solely in a sexual context, 9,10 the analysis of associations between polydrug use and measures of condomless sex was repeated in the subgroup of MSM who reported any anal or vaginal sex in the past 3 months. Within this subgroup, we further examined the adjusted association of specifi c drugs and club drugs with higher-HIV-risk condomless sex with a serodiscordant partner with multivariable modifi ed Poisson models. For each drug, the reference category was no recreational drug use, and the remaining two categories were use of the specifi c drug and use of any other recreational drug. We adjusted all multivariable Poisson models for clinic centre. We used Stata SE (version 12.0) for all statistical analyses.

Role of the funding source
The sponsor of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had fi nal responsibility for the decision to submit for publication.
1435 men who identifi ed as MSM had anal or vaginal sex in the previous 3 months; 1402 (98%) had anal sex with men only, 12 (<1%) had anal or vaginal sex with women only, and 21 (2%) had both anal sex with men and anal or vaginal sex with women.
Although prevalence of drug use was high across most demographic groups, in univariable analysis, recreational drug use was associated with younger age, being employed or unemployed (compared with being retired, a carer, sick or disabled, or a student), rented housing (compared with owning a home, living in temporary accommodation, or being homeless), not identifying with a religion, having an HIV-positive stable partner (compared with an HIV-negative or no stable partner), disclosing HIV status to anyone else, cigarette smoking, and evidence of harmful alcohol drinking (table 1). Recreational drug use was signifi cantly associated with having a viral load greater than 50 copies per mL and with ART use and adherence, being higher in men who were not on ART, and those who were on ART and nonadherent, than in those who were on ART and adherent. Drug use was not signifi cantly associated with ethnic origin, education, fi nancial hardship, time since HIV diagnosis, or CD4 cell count (table 1).
Factors independently associated with recreational drug use (p<0·05) were younger age, not identifying with a religion, having an HIV-positive stable partner, having disclosed HIV status to anyone else, evidence of harmful alcohol drinking, current cigarette smoking, not being on ART, and non-adherence to ART (table 1). ‡Identifi es as belonging to a religion (Christianity, Judaism, Islam, Buddhism, Sikhism, any other). §Evidence of harmful alcohol consumption (that increases the risk of harmful consequences for the user or others) defi ned as a score of 8 or greater on the modifi ed AUDIT-C questionnaire. ¶Non-adherence defi ned as: missed one or more ART dose in the past 2 weeks or missed 2 or more days of ART on one or more occasions in the past 3 months.

Table 1: Association of sociodemographic, HIV-related, and lifesty le factors with recreational and polydrug use in the past 3 months in HIV-diagnosed men who have sex with men
In univariable analysis in 1138 MSM who used at least one drug, use of four or more drugs versus one to three drugs in the past 3 months was associated with: younger age, university education, current employment, not identifying with a religion, having an HIV-positive stable partner, more recent HIV diagnosis, viral load greater than 50 copies per mL, not being on ART, and ART non-adherence (table 1). In multivariable analysis, younger age, university education, not identifying with a religion, and having an HIV-positive stable partner were still signifi cantly associated with use of four or more drugs compared with use of one to three drugs (table 1).
Compared with MSM who did not use drugs in the previous 3 months, those who used drugs were more likely to report any anal or vaginal sex, condomless sex, new STI diagnoses, group sex, searching for sexual partners on the internet in the past 3 months, having ten or more new sexual partners in the past year, and being less likely to use a condom with a casual partner (table 2) With increasing number of recreational drugs used we noted striking increases in the prevalence of condomless sex and all other sexual behaviour outcomes (table 2). These associations were not greatly attenuated after adjustment for statistically signifi cant sociodemographic and HIV-related factors (fi gure 2). We detected a strong association between increasing polydrug use and higher prevalence of all types of condomless sex for the subgroup of 1435 MSM who reported any anal or vaginal sex in the past 3 months (appendix).
We assessed the adjusted association between individual drugs and higher-HIV-risk condomless serodiscordant sex in MSM who reported any anal or vaginal sex in the past 3 months (fi gure 3). Compared with MSM who used no drugs, MSM who used methamphetamine had the highest prevalence ratio of higher-HIV-risk condomless serodiscordant sex, followed by the prevalence ratios for club drugs, erectile dysfunction drugs, nitrites, and cocaine, which were all of a similar magnitude.

Discussion
Our fi ndings show that half of the 2248 HIV-diagnosed MSM surveyed had used recreational drugs in the past 3 months and that about a quarter had used at least three types of drugs during that time period. Drug use and polydrug use were independently associated with younger age, not identifying as religious, and having an HIV-positive stable partner. Drug use was also associated with disclosure of HIV serostatus, harmful alcohol drinking, cigarette smoking, not being on ART,  and non-adherence to ART. We saw strong and consistent associations between increasing numbers of drugs used and increasing prevalence of all indicators of condomless sex (including high HIV-transmission risk condomless sex), group sex, and having multiple new sexual partners. However, the overall prevalence of higher-HIV-risk serodiscordant condomless sex was low (7%).
To the best of our knowledge, ASTRA is the largest questionnaire study of HIV-diagnosed individuals in the UK to date. We think our study population is representative of HIV-diagnosed MSM in the UK because the UK has universal access to health care and 95% of HIV-diagnosed people access specialist services. 11 The response rate (64%) was satisfactory with no signifi cant diff erences in viral load or CD4 cell count between responders and those who did not respond to the questionnaire but who consented to participate. 22 A further strength of the study is the comprehensive information obtained about self-reported polydrug use and recent sexual behaviour. Nevertheless, underreporting of drug use and specifi c sexual behaviours is possible and could have led to underestimation of prevalence.
Previous comparable data for recreational drug use in HIV-diagnosed MSM in the UK derive predominantly from a 2002-03 study that surveyed HIV-negative and HIV-positive MSM at gyms and outpatient clinics in London; 21 prevalence of recreational drug use in the previous year in the HIVdiagnosed outpatient sample (n=388) was 54%. 21 When comparing use of specifi c drugs in the previous 3 months in participants from London clinics in the ASTRA study (N=1527) to drug use that occurred at least once or twice a month in the 2002-03 London HIV outpatient sample, we detected evidence of higher use of methamphetamine (10% vs 4%), cocaine (23% vs 14%), and injecting drug use (4% vs 1%), similar prevalence for use of ketamine (12% vs 13%) and amphetamine (3% vs 2%), and lower prevalence for use

CLS with a seroconcordant partner
No drugs 1 drug 2 drugs 3 drugs 4 drugs ≥5 drugs

CLS with a serodiscordant partner
No drugs 1 drug 2 drugs 3 drugs 4 drugs ≥5 drugs of MDMA (11% vs 17%). 21 This comparison suggests changing patterns of drug use in HIV-diagnosed MSM, with possible increases in methamphetamine and injection drug use. However, comparisons are not straightforward because of the diff erent recall periods for drug use, potential con founding factors in the comparison, and the shortage of data from other comparable studies in the UK.

Higher-HIV-risk CLS-serodiscordant *
There have been few studies of polydrug use in representative samples of HIV-diagnosed MSM in the UK (panel). Findings from the London 2002-03 study showed that more than 90% of the 49 HIV-diagnosed MSM who used methamphetamine during the previous year had used at least one other drug during this time period. Similarly, more than 80% of the 162 cocaine users had used other drugs in the previous year. 21 In our study the high prevalence of polydrug use in the previous 3 months (46% of drug users using three or more drugs and 21% using fi ve or more) is concerning, especially in view of potential drug-drug interactions, such as possible cardiovascular harms from simultaneous use of erectile dysfunction drugs and nitrites, 27 intensifi cation of the toxic eff ects of GHB when consumed with alcohol, 28 and the potential for polydrug use to interfere with the eff ectiveness of antiretroviral drugs. 28 Although we saw that recreational drug use was associated with non-adherence to ART and lower prevalence of suppressed viral load, 87% of HIV-diagnosed MSM on ART who used recreational drugs, and 83% of those on ART who used fi ve or more drugs, had suppressed viral load, showing that recreational drug use is not incompatible with good ART adherence.
In our study, recreational drug use and polydrug use were associated with being sexually active, and more strongly associated with all measures of condomless sex. Although a minority (15%) of HIV-diagnosed MSM reported sero discordant condomless sex, and fewer (7%) fulfi lled our criteria for higher-HIV-risk serodiscordant condom less sex, it was evident that increasing polydrug use was associated with increases in prevalence of serodiscordant condomless sex, and specifi cally higher-HIV-risk condom less sex. In men who reported anal or vaginal sex in the past 3 months, meth amphetamine had the strongest association with higher-HIV-risk sero discordant condom less sex. Findings from previous cross-sectional and longitudinal studies have shown associations between use of meth amphetamine and serodiscordant condomless sex in HIV-diagnosed MSM, 6,21 although none has incorporated measures of viral load. 30 Additionally, results from a US cohort of MSM showed that use of methamphetamine was independently associated with HIV-seroconversion during follow-up. 29 Risk of HIV transmission is low when the HIV-positive partner is adherent to ART with undetectable viral load in the absence of STIs, 31

Figure 3: Adjusted prevalence ratios for the association of use of specifi c drugs in the past 3 months with higher-HIV-risk condomless sex with a serodiscordant partner
Data from 1435 HIV-diagnosed men who have sex with men who reported any anal or vaginal sex in the past 3 months. Adjusted for age group, religion, antiretroviral treatment use and adherence, and clinic. Higher-HIV-risk sex is defi ned as condomless sex plus: not on antiretroviral treatment, latest viral load greater than 50 copies per mL, or new diagnosis of a sexually transmitted infection in the previous 3 months. *Ecstasy, ketamine, mephedrone, and gamma-hydroxybutyrate or gamma-butyrolactone.

Systematic review
We searched PubMed and Embase for articles published in English (Jan 1, 1996, to July 31, 2014) using the following MeSH headings in any fi elds: "recreational drugs" or "street drugs" or "illicit drugs" or "designer drugs" or "controlled substances" or "counterfeit drugs" and "men who have sex with men" or "gay" and "sexual behaviour" and "HIV". We identifi ed cross-sectional studies based on venue-based, online, and clinic-based samples, of recreational drug use in men who have sex with men (MSM) in the USA, Australia, Canada, and Europe, and longitudinal cohort studies of recreational drug use and occurrence of HIV in MSM from the USA. 5,16,30 We identifi ed two small qualitative studies of MSM in the UK who use recreational drugs in a sexual context, irrespective of HIV-serostatus. 9,10 One cross-sectional HIV clinic-based study was done in 2002-03 in HIV-diagnosed individuals in the UK that examined associations with sexual behaviours, such as condomless sex with casual partners and looking for sex on the internet. 21

Interpretation
To the best of our knowledge, ASTRA is the largest questionnaire study of HIVdiagnosed individuals in the UK, accounting for about 5% of all HIV-diagnosed MSM in the UK. Our study includes comprehensive information about the extent of polydrug use, which has not been reported before in the UK. We measured condomless sex with HIV-seroconcordant, HIV-serodiscordant or unknown status partners, and condomless sex with serodiscordant partners accounting for viral load, which has not previously been examined. Our fi ndings show that recreational drug use and polydrug use are prevalent, and strongly associated with condomless sex and multiple new sexual partners among HIV-diagnosed MSM. These fi ndings draw attention to the importance of tailored harm reduction support for MSM who use recreational drugs and to the need to address the prevention of HIV and sexually transmitted infection in this group.
of transmission during anal sex in MSM. 32 However, such condomless sex, as well as condomless sex with partners known to be HIV-positive, does present a risk of transmission of other STIs including hepatitis C infection, syphilis, and gonorrhoea. Additionally, since 2004, the UK has an ongoing lymphogranuloma venereum epidemic in MSM, of which most cases are in HIV-diagnosed MSM, among whom reported levels of drug use (particularly methamphetamine and mephedrone use) are high. 33 The diffi culties in attributing a causal relation to the associations between recreational drug use and higherrisk sexual behaviours in cross-sectional studies have been well documented. 34 However, irrespective of causal attributions, our fi ndings show that polydrug use and condomless sex are inextricably linked in HIVdiagnosed MSM in the UK, and that polydrug users are likely to be a group at especially high risk for transmission of HIV and other STIs. A need exists for longitudinal, episode-level studies focusing on drug use during episodes of condomless sex in the same individual, which could provide important information about temporality and causality.
Improved understanding of the underlying drivers of polydrug use in HIV-diagnosed MSM is needed in order to reduce health harms. Cross-agency collaboration between HIV treatment and substance misuse services might be benefi cial in providing tailored, judgment-free harm reduction advice and support to HIV-diagnosed MSM who use recreational drugs, and in addressing HIV and STI prevention issues in this group. Peer-led interventions might also be productive in outreach services for HIV-diagnosed men who are polydrug users and have multiple sexual partners. National STI and HIV prevention strategies should address recreational drug use.