Content uploaded by Matthew Hibbert
Author content
All content in this area was uploaded by Matthew Hibbert on Oct 11, 2021
Content may be subject to copyright.
A narrative systematic review of sexualised drug use and sexual health outcomes among
LGBT people
Abstract
Introduction: Chemsex is a specific form of sexualised drug use (SDU) that is an emerging
public health issue among men who have sex with men (MSM). Although the recent focus on
chemsex is a reflection of the associated harms it is important to understand SDU more
broadly and its associations with risk behaviours. Additionally, some of the reasons suggested
for MSM engagement in SDU are also likely to apply to women who have sex with women
(WSW) and trans people. The aim of this review was to investigate SDU, including chemsex,
among lesbian, gay, bisexual and trans (LGBT) people internationally in relation to sexual
health outcomes (HIV status, STI diagnosis, condom use).
Methods: Papers that were published between January 2010 and June 2020 reporting SDU in
MSM, WSW, or trans people were identified through Medline, PsycINFO, CINAHL Plus and
Web of Science. Results were synthesised using a narrative approach.
Results: The search identified 2,710 publications, of which 75 were included in the final
synthesis. The majority of studies measured SDU among MSM (n=71), and four studies
measured SDU among trans people. Research into SDU had been conducted in 55 countries
and 32 countries had recorded the use of a chemsex drug among MSM, although the drugs
used to define chemsex varied. Among studies that researched MSM, SDU was most
commonly investigated in relation to condomless anal intercourse (n=42), followed by HIV
prevalence (n=35), and then STI diagnoses (n=27). Drug use was generally associated with
sexual health outcomes, but particularly in chemsex studies.
1
Conclusions: SDU research is lacking among WSW and trans people, despite trans women
having a high HIV prevalence. Among MSM, most drugs were associated with sexual health
outcomes, and therefore it is important to include both chemsex drugs and other drugs in
SDU research.
Keywords: LGBT people; sexualised drug use; chemsex; systematic review; sexual health
2
Introduction
Sexualised drug use (SDU) is a term used to refer to sexual activities whilst under the
influence of a wide range of drugs and substances, such as cannabis, amyl nitrates (poppers),
and crystal methamphetamine and has been a topic of research among men who have sex
with men (MSM) for some time (Bourne, 2012; Leigh & Stall, 1993). Chemsex (sometimes
referred to as ‘party and play’) is a particular form of SDU whereby men engage in sex with
other men for long periods of time with multiple sexual partners, typically taking one or more
of crystal methamphetamine, γ@hydroxybutyrate/ γ@butyrolactone (GHB/GBL), methedrone,
cocaine and/or ketamine immediately before or during sex to facilitate and enhance the
sexual experience (Bourne, Reid, Hickson, Torres Rueda, & Weatherburn, 2014). Chemsex
has become a public health issue over the last decade. Clinicians (Stuart, 2013) and men who
engage in chemsex (Ahmed et al., 2016) report that this may be due to an increase in the
number of people engaging in this behaviour, with geospatial networking apps and online
sites used to meet sexual partners likely to have enabled this increase (Ahmed et al., 2016;
Stuart, 2013).
Chemsex has been associated with injecting drug use and sexual risk behaviours such
as condomless anal intercourse (CAI) and a greater number of CAI partners (Bourne et al.,
2014; Glynn et al., 2018; Hegazi et al., 2017), as well as with a greater likelihood of sexually
transmitted infection (STI) diagnoses and human immunodeficiency virus (HIV) (Gilbart et
al., 2015; Glynn et al., 2018; Hegazi et al., 2017; Hibbert, Brett, Porcellato, & Hope, 2019a).
The broader sexualised use of drugs among MSM has been a topic of research for much
longer, due to the potential for HIV acquisition from both injecting drug use and CAI
sometimes facilitated by drug use (Halkitis, Parsons, & Stirratt, 2001; Mattison, Ross,
Wolfson, Franklin, & HNRC Group, 2001; Stall & Purcell, 2000). Although there is no
agreed definition of chemsex or even of drugs associated with chemsex, people may engage
3
in a variety of forms of SDU that encompasses what may be categorised as chemsex as well
as other forms of SDU. However, due to the recent focus on chemsex, SDU more broadly has
received less attention, even though it incorporates a much wider range of drug use
behaviours and therefore may be more common.
A narrative systematic review of international quantitative and qualitative chemsex
literature found that chemsex among MSM (defined as taking methamphetamine,
mephedrone, GHB/GBL, cocaine or ketamine before or during sex) was associated with high
sexual risk and HIV prevalence (Maxwell, Shahmanesh, & Gafos, 2019). Additionally, a
systematic review investigating international prevalence of SDU among MSM found the
most commonly reported sexualised drugs were methamphetamines, GHB/GBL and amyl
nitrates (poppers), and that SDU was associated with increased sexual risk (Tomkins, George,
& Kliner, 2019). A literature review investigating prevalence of both chemsex (defined as
taking methamphetamine, GHB/GBL and mephedrone before or during sex) and SDU
(defined as use of illicit drugs before or during sexual activity) in the United Kingdom (UK)
found the prevalence of both varied across studies (Edmundson et al., 2018). The reviews
identify that both SDU and chemsex are associated with sexual risk, and that the definitions
of chemsex and SDU vary between studies. A recent UK study found that those MSM who
engaged in SDU have a greater likelihood of a STI diagnosis and more anal intercourse
partners than those who had not engaged in SDU, and among those who had engaged in
SDU, those who engaged in chemsex (defined as the use of crystal methamphetamine, GHB/
GBL, ketamine, mephedrone) had a greater likelihood of an STI diagnosis and more anal
intercourse partners (Hibbert, Brett, Porcellato, & Hope, 2019b). Therefore, those engaging
in chemsex may be engaging in greater sexual risk behaviours than those who engage in other
forms of SDU.
4
Research investigating SDU among MSM has mostly been conducted in Western
countries, and the term chemsex is typically used in a Western context (Bourne &
Weatherburn, 2017), but SDU, including the sexualised use of drugs associated with
chemsex, has also been observed internationally (Bourne, 2012; Maxwell et al., 2019;
Tomkins et al., 2019). Due to the nature of researching SDU, studies tend to be cross@
sectional, making causation hard to infer. Causation is also hard to infer when researching
SDU due to the large number of potentially confounding factors that could also impact on
sexual health outcomes (e.g. relationship status, number of sexual partners). Despite this,
Leigh and Stall (1993) categorised the possible ways of measuring and analysing SDU into
global associations, situational associations, and event@level associations. Global association
is where general drug use is measured over a specific period (e.g. in the past 12 months), and
sexual behaviour is also measured over a specific period, and an analysis between the two is
conducted. Situational association is where the drug use is measured in relation to sex over a
specific period, and sexual behaviour is also measured over a specific period, and an analysis
between the two is conducted. Event@level associations are where drug use and sexual
behaviour are asked about for a specific sexual event (e.g. the last time you had sex whilst
using a drug, did you use a condom?). This is particularly important when investigating SDU,
including chemsex, as drug use and sex are both linked, so a valid measurement needs to
account for this (by using situational or event@level associations) rather than to explore global
associations.
Some of the factors that have been suggested as motivations for MSM engaging in
SDU, including chemsex, such as internalised homophobia and HIV stigma (Weatherburn,
Hickson, Reid, Torres@Rueda, & Bourne, 2017), may apply to SDU among other lesbian, gay,
bisexual and transgender (LGBT) people. This is particularly pertinent for trans women who
have a high global estimate of HIV prevalence (Baral et al., 2013). However, less is known
5
about SDU among trans women and if it is associated with the same sexual risk as MSM.
Additionally, SDU has been observed among women who have sex with women (WSW)
(Hibbert, Porcellato, Brett, & Hope, 2019), but comparatively little research has been
conducted in this group.
Therefore, the primary aim of this review was to:
Investigate the associations between SDU and specific health outcomes (HIV status,
STI diagnoses and CAI) among LGBT people.
Secondary aims of this review were to:
Investigate how representative research into SDU in relation to health outcomes is of
the whole LGBT population.
Analyse what methods are used to explore the relationship between SDU and health
outcomes (global association, situational association or event@level association).
Assess which countries have reported SDU among LGBT people, and in particular,
which countries have reported the sexualised use of a drug associated with chemsex.
Method
The systematic review was designed and reported following the PRISMA guidelines
(Moher, Liberati, Tetzlaff, & Altman, 2010), with the protocol registered at PROSPERO
International Register of Systematic Reviews prior to commencing the review (ID
CRD42018084366). The PECO framework (Population, Exposure, Comparison, Outcome)
(Methley, Campbell, Chew@Graham, McNally, & Cheraghi@Sohi, 2014) was used to form the
search strategy where the population was LGBT people (MSM, WSW, and trans people);
exposure was SDU; comparison was between those engaging in SDU and those who were
not; and the outcome was HIV, STI diagnoses, or CAI. These outcomes were decided upon
6
due to preliminary searches identifying these outcomes as commonly measured in relation to
SDU and chemsex. The analysis between SDU and HIV, STI diagnoses, or CAI was
classified as global, situational, or event level associations. Due to the difficult nature of
attributing HIV and STI transmission to specific sexual events, only the measurement of CAI
could be classified as an event@level association. Studies that measured HIV and STIs by self@
report and/or laboratory tests were included in the review.
Suitable search terms were obtained from systematic reviews on similar topics (Choi,
Wong, & Fong, 2017; Vosburgh, Mansergh, Sullivan, & Purcell, 2012). A preliminary search
was then conducted using these terms on MEDLINE, with relevant articles retrieved to
identify additional search terms. Search terms were grouped into three concepts: “LGBT
terms”, “Drug terms”, and “Sex terms” (Table 1),so that searches used the string: “LGBT
terms” AND “Drug terms” AND “Sex terms”. The search string was used to search
MEDLINE, PsycINFO, CINAHL Plus, and Web of Science (EBSCO MEDLINE from 1879
to 30th June 2020, ProQuest PsycINFO from 1806 to 30th June 2020, EBSCO CINAHL Plus
from 1981 to 30th June 2020, Web of Science Core Collection from 1900 to 30th June 2020).
Where studies were not published in English, an attempt to find a translation was made. A
period limit of 1st January 2010 to 30th June 2020 (inclusive) was imposed due to the end
date of a previous systematic review on a similar topic (Vosburgh et al., 2012), although that
systematic review was specific to MSM engaging in SDU and event@level condom use. A
limit was also set on the period of data collection (January 2010@June 2020) to ensure that the
review represented recent patterns of SDU. An attempt to find grey literature from relevant
community organisations and public health organisations was made, but no reports provided
sufficient detail to be included.
Chemsex drugs were defined as the ‘4 chems’ (crystal methamphetamine, GHB/GBL,
ketamine and mephedrone) as in Schmidt et al. (2016), and are commonly accepted as drugs
7
associated with chemsex (Bourne et al., 2014). SDU and their health related outcomes were
grouped into three categories: global association, situational association, and event@level
associations (Leigh & Stall, 1993).
Four stages were used to identify studies: identification, screening, eligibility and
inclusion (Moher et al., 2010). A data extraction form with quality assessment was adapted
from The Cochrane Public Health Group Data Extraction and Assessment Template, The
Effective Public Health Practice Project (EPHPP) Quality Assessment Tool for Quantitative
Studies, and the Center for Evidence@Based Management (CEMa) critical appraisal checklist
for surveys. This form was created using online survey software Qualtrics
(www.qualtrics.com). Cross@tabulated data on drug use and outcomes (HIV status, STI
diagnoses, condom use) were extracted, and where underlying data were not available, the
unadjusted odds ratios were extracted. There was not enough homogeneity in the data in
terms of time duration of drug use or sexual risk factor (i.e. ever, the last 12 months, the last
three months), as well as in terms of the populations studied (i.e. MSM living with HIV only,
sex workers, young MSM, Black MSM only) for meta@analyses to be conducted.
Additionally, it was not possible to control for confounding variables (e.g. age, number of
sexual partners, sexuality, relationship status) due to the variability in the confounding
variables reported in the studies. Screening and eligibility was conducted by two researchers
independently (MH and AH) and a third researcher (VD) was used for any disagreements.
Data extraction was completed by one researcher (MH), and then checked by a second
researcher (AH). Quality assessment was conducted by two researchers independently (MH
and AH), and a third researcher was used for any disagreements (VD).
*Table 1 about here*
8
Inclusion criteria:
1. A measure of drug use and sexual health risk (HIV status, STI diagnoses, CAI) within
the same population.
2. An association analysis conducted between the drug behaviour and the sexual
behaviour and/or health risk.
3. Population studied included a sexual and/or gender minority population.
4. Studies published and data collected in the date range January 2010 to June 2020
(inclusive).
Exclusion criteria:
1. Articles not published in English and no translation available.
2. Studies including children (aged equal to or less than 15 years).
3. Studies that are not relevant to the research question (e.g. heterosexual populations
only, comparisons with heterosexual people, medical drug trial studies).
4. Qualitative research.
Results
The search yielded 2,710 unique citations, of which 1,658 were excluded during title
and abstract screening and 977 during full@text review, leaving 75 studies eligible for data
extraction (Figure 1). The majority of studies were cross@sectional (n=67), six studies were
cohort studies, and two were case@control studies. A list of included studies can be found in
Supplementary Material 1.
*Figure 1 about here*
9
The 75 included studies spanned across 55 countries, of which 71 researched MSM and 4
researched trans women (Supplementary Material 2). No studies examining SDU among
WSW or trans men/non@binary people met the inclusion/exclusion criteria. The countries of
included studies in which SDU had been studied among MSM or trans women or both are
shown in Figure 2, as well as the countries where a study had examined the use of one or
more chemsex associated drugs (crystal methamphetamine, GHB/GBL, ketamine, or
mephedrone) among MSM.
*Figure 2 about here*
Men who have sex with men
Over one quarter of studies among MSM were conducted in the United States of
America (USA) (n=20), and around a quarter were conducted in the UK (n=16). The most
commonly studied drug among MSM was cannabis (n=34), followed by amyl
nitrates/poppers (n=29), crystal methamphetamine (n=18), erectile dysfunction drugs (EDD)
(n=15), cocaine (n=13), ecstasy (n=10), GHB/GBL (n=11), and ketamine (n=8). Less
common drugs were mephedrone (n=3), heroin (n=2), amphetamine (n=1), and crack cocaine
(n=1). The time period for measuring drug use ranged from the last anal sex event (n=6) to
lifetime use (n=3), with the most common recall period being six months (n=20).
Around one third of studies (n=23) grouped drugs into chemsex/party drugs. The
specific drugs grouped as chemsex/party drugs varied considerably, but GHB/GBL were
included in all chemsex groups, and crystal methamphetamine was included in 21 out of 23
10
chemsex groups (Table 2). Fourteen of the studies used situational association analyses to
investigate the sexual health outcome and the remaining nine used global association
analyses. At least one of the drugs associated with chemsex had been investigated in the
majority of countries (n=32/54, Figure 2).
*Table 2 about here*
HIV prevalence was the outcome examined in 35 studies, 27 studies examined STI
incidence, and CAI was the most common health outcome examined (n=42). . Table 3
displays the breakdown of outcomes studied in relation to specific drugs and how many
studies found a significant bivariate association between drug use and the outcome
investigated.
*Table 3 about here*
HIV prevalence
Among the 35 studies that investigated HIV prevalence, ten measured HIV status
using a laboratory test, the remaining 25 asked participants to self@report their HIV status.
The majority of studies conducted global association analyses (n=26) and nine conducted
situational association analyses. The most common drug category investigated in relation to
HIV prevalence was poppers use (n=14) with around 70% of studies finding a bivariate
association between poppers use and HIV prevalence. Chemsex drugs grouped were
investigated in relation to HIV prevalence in thirteen studies and 77% of studies found
11
bivariate associations between chemsex drug use and HIV prevalence. For chemsex drugs
that were investigated independently, four out of the seven studies investigating crystal
methamphetamine use and HIV prevalence found a bivariate association. Four studies were
found that investigated either GHB/GBL or ketamine and two found bivariate associations
with HIV prevalence (50%). Two studies investigated mephedrone and HIV prevalence and
both found a bivariate association, although both studies investigated among MSM who use
drugs only and three out of the four studies that investigated EDDs found a bivariate
association with HIV prevalence.
STI diagnoses
Among the 27 studies investigating STI diagnoses in relation to drug use, 16
conducted global association analyses, 10 conducted situational association analyses and one
study conducted both global and situational association analyses. Most studies grouped STIs
for their investigations (n=17). Seven of these groups investigated associations between drug
use and bacterial STIs only (chlamydia, gonorrhoea, lymphogranuloma venereum (LGV),
syphilis), whilst others included herpes, genital warts, and blood borne viruses such as
hepatitis A, B, and C as well as newly acquired HIV (See Supplementary Material 2). Over
half (n=15) of the studies measured STI diagnoses using laboratory tests, whilst the
remaining 12 studies used self@report methods. The recall period for self@reported STI
diagnoses was most commonly 12 months (n=8), whilst two studies had a recall period of
three months, one study had a recall period of six months, and one study did not report the
recall period.
The most commonly researched drug in relation to STI diagnoses were chemsex drugs
grouped (n=14), and in the majority of studies (n=11) this was in relation to grouped STI
diagnoses. One study each investigated chemsex drugs in relation to shigella, hepatitis C and
12
gonorrhoea, and all bivariate associations between chemsex drugs and STI diagnoses were
significant. Poppers were the second most commonly researched in relation to STI diagnoses
(n=11). Six studies investigated poppers use in relation to syphilis diagnoses, all of which
were conducted in China, with two of these studies finding bivariate associations. All four
studies that conducted bivariate association analyses between poppers use and grouped STI
diagnoses found associations, and one study investigated poppers use in relation to Shigella
diagnoses and found a bivariate association.
Condomless anal intercourse (CAI)
Among the 42 studies investigating CAI, 22 studies conducted global association
analyses, 14 conducted situational association analyses, five conducted event@level
association analyses, and one study conducted both global and situational association
analyses depending on the drug of interest. The recall period for CAI ranged from event@
based (n=6) to 12 months (n=7), with the most common recall period being three months
(n=13). The most commonly researched drug in relation to global and situational analyses for
CAI was cannabis (n=19), with around 40% of studies (n=8) finding a bivariate association
between cannabis use and CAI. All 11 studies that investigated chemsex drugs grouped found
bivariate associations when analysing CAI and over three quarters of studies (n=13/16) that
investigated poppers use and CAI found a bivariate association.
Among the five studies that used event@level associations, all three studies
investigating crystal methamphetamine, EDDs, GHB/GBL, and poppers found bivariate
associations between drug use and CAI. One study found bivariate associations between
cannabis use, ecstasy use and CAI among HIV positive partners only, whilst the remaining
three studies that investigated event@level associations between cannabis use, ecstasy use and
13
CAI found no association. One study investigated each cocaine and ketamine use in relation
to CAI and both found no association.
Trans women
Among the four studies that researched trans women, a range of drugs were
investigated (cocaine, crack cocaine, crystal methamphetamine, heroin, and poppers). Three
studies conducted global association analyses and one study conducted a combination of
global and situational analyses. Two studies were conducted in the USA, one in Brazil, and
one in Vietnam. A bivariate association was found between crystal methamphetamine (n=1)
and condomless sex. One study investigated a possible association between poppers use and
condomless sex, but found no association. One study found a situational association between
cocaine and HIV status, although two studies did not find this association when conducting
global association analyses. One study found an association between methamphetamine use
and HIV status. No association was found between heroin use (n=1) and HIV status or
syphilis diagnosis.
Quality assessment of the included studies
The majority of studies had an overall rating of moderate (n=31/75, 41%), 28 studies
(33%) were rated as weak, and 16 (21%) were rated as strong. The weakest sections tended to
be the reporting of withdrawals and dropouts, where 34 studies (45%) were rated as weak,
and confounders, where 20 studies (27%) were rated as weak.
14
Discussion
This systematic review firstly assessed the extent to which SDU has been studied
among LGBT people in relation to sexual health outcomes, and it found that the vast majority
of research has been conducted among MSM. A smaller number of studies had been
conducted among trans women; however, no studies were found that reported on SDU in
relation to sexual health outcomes among trans men or WSW. Whilst a few studies were
found among trans women, due to the potentially high risk of HIV among trans women
(Baral et al., 2013), further studies are needed to explore SDU and its related sexual and
health implications among trans women.
It should be noted that some studies among WSW were found, but they were not
included in this review because they compared WSW with heterosexual women, and
therefore data were not available for health outcomes exclusively for WSW. These few
studies identified that WSW may be more likely to use ketamine (Heinsbroek, Glass,
Edmundson, Hope, & Desai, 2018), as well as cannabis and cocaine (Bauer, Jairam, &
Baidoobonso, 2010) compared to heterosexual women, but drug use was not measured in a
sexual context. One study did find that lesbian and bisexual women were more likely to
engage in SDU (Estrich, Gratzer, & Hotton, 2014), but data were not available exclusively
for WSW with regards to sexual risk and SDU. Recent research has indicated an association
between SDU among WSW and greater sexual risk (Hibbert, Porcellato, et al., 2019), but
further event@level research among WSW is warranted. Narrowing the search criteria to
sexual health outcomes may have limited the number of possible studies found in relation to
WSW, as the reasons for engagement and effects on sexual health may be different for WSW
(Hibbert, Porcellato, et al., 2019). However, due to bisexual women possibly being more
likely to engage in SDU (Estrich et al., 2014; Hibbert, Porcellato, et al., 2019), research
regarding sexual health behaviours like condom use and STI diagnoses may be warranted.
15
Whilst there was some research regarding SDU among trans women, further research is
needed due to the suspected high@risk of HIV transmission. Additionally, no studies indicated
they included trans men and non@binary people. Where trans men do identify as MSM, they
are most often not included in analyses among men. Therefore, further research is needed
among trans people in general, to understand if SDU exists, and if so, whether it is associated
with sexual risk.
Associations between SDU among MSM and/or trans women have been studied in 55
countries, but due to the inclusion and exclusion criteria used in this review, not all countries
which have examined SDU among LGBT people may have been identified. Additionally, the
use of at least one of the drugs associated with chemsex among MSM has been observed in
32 of these countries spanning North America, Europe, Asia, and Australasia. Although
chemsex was first documented within the UK and has been researched in other Western
countries (Bourne & Weatherburn, 2017), this behaviour has been observed internationally,
and therefore more international research is needed, particularly in countries with high
prevalence of HIV among MSM. Similar to a literature review of SDU and chemsex in the
UK (Edmundson et al., 2018), it was found that the definition of chemsex varied greatly, but
GHB/GBL was included in all chemsex definitions. This may be because the drugs used for
chemsex differ internationally, or that research had been conducted before a definition which
drugs used specifically for sex, constituted as chemsex. A consensus of what drugs constitute
as chemsex may be hard to reach due to emerging new drugs, local availability of specific
drugs, or personal preferences for the type of drugs used for chemsex. Therefore, an
international definition of what drugs constitute as chemsex may not be appropriate and
instead more local definitions may be more suitable. Although the lack of an international
definition may limit cross@cultural comparisons, because of the sexual risk associated with
chemsex, defining chemsex with regards to sexual behaviour may be equally or if not more
16
important, and trying to commonly define chemsex as a particular use of drugs may be too
simplistic. However, it is useful to see which drugs are common internationally, so harm
reduction and drug safety information can be shared across countries.
It is of note that nearly all drugs were associated with greater sexual risk, regardless of
the drug or outcome in question, similar to previous systematic reviews regarding chemsex,
SDU and MSM (Maxwell et al., 2019; Tomkins et al., 2019), but causation cannot be inferred
from these analyses. For some of the more commonly used, and more socially accepted
substances, such as poppers and cannabis, the associations found may in part be due to people
taking these substances also being more likely to use other substances (which they may not
always disclose). It is unclear the influence that polydrug use may have on these findings. For
example, whether individuals who use multiple drugs during the same sexual encounter are
considered to engage in greater risk taking. When considering global associations, individuals
may use multiple drugs in a variety of different contexts, some of which are specific to a
sexual context. However, in these situations, drug use outside a sexual context could be
associated with a sexual health outcome via proxy. Therefore, there is a need to move away
from global associations when investigating SDU and sexual health associations.
Certain patterns of drug use, such as chemsex, may be associated with HIV
prevalence, STI diagnoses and CAI more than other patterns of drug use, which has been
suggested by previous research (Hibbert, Brett, et al., 2019a). The variation found in this
review regarding the definition of what drugs constitute as chemsex may be because sex
under the influence of the drugs used in chemsex may lower inhibitions and therefore impact
on behaviour to a greater extent than other types of SDU, or social norms associated with
chemsex may influence risk taking. Additionally, it could be that grouping drugs creates a
more powerful analysis due to a greater number of observations included, and therefore this
is why chemsex appears to be associated with greater risk.
17
There were similar associations between event@level analyses and global and
situational associations for condom use, but a large number of studies relied on global
associations of drug use and health outcomes, even when aiming to research chemsex, which
is by definition use in a sexual context. Therefore, if future research is aiming to investigate
SDU, situational and event@level analyses should be used for a potentially more accurate
measurement. Another limitation identified in the research was the variability in the recall
period for reporting drug use and sexual health behaviours. Whilst lifetime use of drugs may
be important for drugs that are typically injected when investigating blood borne viruses, for
more recent sexual behaviours, the usefulness of lifetime drug use is questionable. Therefore,
studies should aim to have more recent recall periods for both drug use and sexual
behaviours. Due to the nature of researching SDU being mostly cross@sectional, causation
cannot be inferred, regardless of the measurement method chosen. It is possible that other
factors influence associations between drug use and sexual risk behaviours, for example,
those who take drugs and have CAI may just be less risk averse.
Strengths and limitations
A limitation of this review is that there may be a publication bias in the data,
suggested by most studies finding an association with the health outcome researched. An
attempt was made to find grey literature on the topic, however no reports or publications were
found where information had not already been published in peer@review journals, or that met
the inclusion/exclusion criteria. Additionally, it was not possible to control for confounding
variables that may influence drug use and HIV, STI diagnoses and condom use, such as age
and sexual identity, due to the heterogeneity among control variables in multivariable
analyses. Collating data is also difficult due to different window periods of measurement (e.g.
three months/six months/twelve months), and variability in the grouping of drugs associated
with sex. In terms of quality, one third of studies were given an overall quality assessment
18
rating of weak. Due to the various sub@categories in this review, it was deemed inappropriate
to exclude studies rated as weak, which may have impacted the findings.
This review focused specifically on HIV, STI diagnoses, and CAI due to these
outcomes being commonly researched in relation to SDU. However, this does ignore other
sexual health factors that may be related to SDU such as pre@ and post@exposure prophylaxis
(PrEP/PEP) use, as well as possible psychological associations. Despite this, the finding that
SDU among MSM was associated with HIV prevalence, STI diagnoses and CAI is still
important when considering service delivery, as well as harm reduction services, due to
potentially confounding factors a person may experience (i.e. drug harms, living with HIV,
greater sexual risk taking).
Conclusion
For the majority of drugs examined, drug use appears to be associated with living
with HIV, STI diagnoses, and CAI among MSM. However, the measurement of SDU often
relied on global associations between drug use and risk, so may be subject to
misclassification bias. Therefore, more accurate measurements of SDU, such as situational or
event@level analyses, should be used. Definitions of what constitutes chemsex drugs varied
across studies, making conclusions with regards to associated risks with chemsex difficult.
The definition of what constitutes chemsex drugs may be even more difficult when
considering the behaviour internationally, as the availability of certain drugs will differ across
countries, depending upon legal categorisation and common illicit drug markets. Further
research is needed regarding SDU among WSW and trans people to assess the occurrence of
SDU and any possible impact on sexual risk.
19
References
!"#$%#"&''()*#
+,'# -#"-# #(("".+/ "./,#((0
'#1##&
#234(5#
*-'#%'#-'6176%8)*09+,#-:
#;<="(:#2"0"> ?:&-0""
+,'&+,&
*7@@)*###"#'A+,'.--"!*?#"'"
B"#".-&<>C#$-'."2B'D>#-'?0
#;$-07#: +,&
#23E1F;
*#+,!""#!"$%"&
?;#G5233"(" #(!3C-"395HHA'AH5;2
*# !"#$%#"')+,"'"$
())""!*
!"&?;#G5"233:::-(#?!3""3;-3C-"3""& ".&
"0&C-&&5#5;2
*#)+,'" "#(:#?".:25G"
#>?>#"5 "?>#?-#5
++,&
9#8#(@I)$#(I%+,%"#;"# -:#!(55- >#"#;
"5#""# ".-"!"-"(0".-"(
5#5->#"2"0"> ?:,'-++,&
#23
8"#9"#!87-""#5=A#A)"A+,
'.-"("B(#+B,2?:#;-"
.//&
8" 9776*)#G#1+,J "'.--"!*?#"
'" B"#(#0'.-<>029 (#&
0+,#-K
7-=1'"<@!"9$(#A7#AE-?<("7+,'.
(""5#55- >#"2C(";#"" ?:"::#
?".:(#":'(--L.8(--"
++,&
7-0*0DEM''-09#A89-!'+,9".
"!?#"".--0"G; >#"#(:#?".:
-<-/-&
-!>"D"#"@%)'>GA@+,#-5 2 0"-5
("->##<="""##((0.10+,
&
(61A@G!7''"9G"!A+,
9". 02".-""" "(0".-#:#?".
:G(".-- - ".2,-+,&
#23
"#!87-""8"#9#5=)"A+,G"#;4 >(#&
4 >(("0".-?#5#5-:#4 ("G("? "
8(--"D#<-&".
#234(5#
20
A*G98# --#1)#5=+,"0 #"# -".-
"> """# :".-"(" ".#(:#?".
:+A'A,B+/+,&
#23"."&&
A*G98# --#1)#5=+,"0 #"# -".-
"> """# :".-"(" ".#(:#?".
:+A'A,B+/+,"."&&
A# --#1*G98)#5=+,""# >#":("
".-"("#(:#:#?".::#+',B2$("
;#17*%'.1;"0-"'?0 0&
#234(5#
1(*9)'--+,'" ""!0".-?#;#.5#"#<=<"""
##-#(05>#5?>#","0+,&
AN"#A#""A#-;"#%$!-)D97#5+,9 50
G -("";".(0),)+&,
#3'&+,&
A.:--''"A)7;#"A+,9".?#"#(:#?".
:2"0"> ?:#;-3&
#234(5#
A-0A95--'9:&79A D--0)9(&'#'+,<9E<9E'
'<82 #5"#"0#;"5 C 0"">?0" ##-";#
K->?"0"> ?:"45'1"#60#2%D
3"&&&
A#1>%6-J@)-7+,;5#>(";#"0">
?:"&-0""2<'A".
789+,&
' @*#A "B) !"#$+,<-- ("
#((0".- >"2$(";#8#5A'A<'?0
+8A<',".ê(5#
'--) --+,<:(85 "2?:#;" #'" B"
#(A#?'.:A<"9# >##<'85 ,
4"#0+,è
'+,'.-"("0A'A2 !(# """5#"1:
;+,&
%#!"7#()-A+,'.-"(!(#(:#?".:
2"0"> ?:%#)1"++,&
#23
=#"(A"(7'--?') --+,?:#;-#
?&-?-"" "".-"!?##(:#?".:
,4"#3+,"&&&
!"#$%#"&')*#+,A#>?>#"?-"
""# : #(".-- ("+M ".M,#((0'#1##2
C(";#K->?"0++,&
#23"."&&
21
Table 1. Results generated from each search term used for each database, June 2020.
Database Keywords Articles retrieved
LGBT terms Drug terms Sex terms
Medline
Homosexuality (MH) Substance@related disorders (MH) "Sexual health"
1,858
Homosexuality, female (MH) "Substance use" Reproductive health (MH)
Homosexuality, male (MH) Alcohol drinking (MH) "Sexual behavior"
Homosexual* Alcohol "Sexual behaviour"
Gay "Drug use" Sexual behavior (MH)
Lesbian* Chemsex "Sexual risk"
Bisexual* "Party and play" Risk@taking (MH)
Transexual* Marijuana Unsafe sex (MH)
Transsexual* GBL "Unsafe sex"
Transgender* GHB
Trans Ecstasy
Transgender persons (MH) Cocaine
Genderqueer Crack
"Non binary" Methamphetamine
"Men who have sex with men" Methadone
"Sexual minorit*" Poppers
Sexual minorities (MH) "Amyl nitrate*"
LGBT* Ketamine
Viagra
"Erectile dysfunction drug*"
"Sildenafil Citrate" (MH)
PsycINFO Homosexuality (SH) Substance@related disorders (SH) "Sexual health" 1,445
Homosexuality, female (SH) "Substance use" Reproductive health (SH)
Homosexuality, male (SH) Alcohol drinking (SH) "Sexual behavior"
Homosexual* Alcohol "Sexual behaviour"
Gay "Drug use" Sexual behavior (SH)
Lesbian* Chemsex "Sexual risk"
Bisexual* "Party and play" Risk@taking (SH)
Transexual* Marijuana Unsafe sex (SH)
Transsexual* GBL "Unsafe sex"
Transgender* GHB
Trans Ecstasy
Transgender persons (SH) Cocaine
Genderqueer Crack
"Non binary" Methamphetamine
"Men who have sex with men" Methadone
"Sexual minorit*" Poppers
Sexual minorities (SH) "Amyl nitrate*"
LGBT* Ketamine
Viagra
22
"Erectile dysfunction drug*"
"Sildenafil Citrate" (SH)
CINAHL Plus
Homosexuality (MH) Substance@related disorders (MH) "Sexual health"
727
Homosexuality, female (MH) "Substance use" Reproductive health (MH)
Homosexuality, male (MH) Alcohol drinking (MH) "Sexual behavior"
Homosexual* Alcohol "Sexual behaviour"
Gay "Drug use" Sexual behavior (MH)
Lesbian* Chemsex "Sexual risk"
Bisexual* "Party and play" Risk@taking (MH)
Transexual* Marijuana Unsafe sex (MH)
Transsexual* GBL "Unsafe sex"
Transgender* GHB
Trans Ecstasy
Transgender persons (MH) Cocaine
Genderqueer Crack
"Non binary" Methamphetamine
"Men who have sex with men" Methadone
"Sexual minorit*" Poppers
Sexual minorities (MH) "Amyl nitrate*"
LGBT* Ketamine
Viagra
"Erectile dysfunction drug*"
"Sildenafil Citrate" (MH)
Web of Science
Homosexual* Substance@related disorders "Sexual health"
1,357
Gay "Substance use" "Sexual behavior"
Lesbian* Alcohol "Sexual behaviour"
Bisexual* "Drug use" "Sexual risk"
Transexual* Chemsex "Risk@taking"
Transsexual* "Party and play" "Unsafe sex"
Transgender* Marijuana
Trans GBL
Genderqueer GHB
"Non binary" Ecstasy
"Men who have sex with men" Cocaine
"Sexual minorit*" Crack
LGBT* Methamphetamine
Methadone
Poppers
"amyl nitrate*"
Ketamine
Viagra
"Erectile dysfunction drug*"
"Sildenafil Citrate"
MH – Medical Subject Heading (MeSH). SH – Subject Heading
23
Table 2. Summary of drugs included in chemsex definitions
Drug or drugs included in chemsex definitions N (n=23)
Crystal methamphetamine 21
GHB/GBL 23
Ketamine 12
Mephedrone 20
Grouped
Crystal methamphetamine, GHB/GBL, mephedrone 20
Crystal methamphetamine, GHB/GBL, ketamine, mephedrone 11
24
Table 3. Bivariate associations found in studies investigating drug use in relation to HIV prevalence, STI diagnoses, and condom use among MSM.
Drugs investigated
HIV prevalence (n=35) STI diagnoses (n=27)* Condomless anal intercourse (n=42)*
Global (n=26) Situational (n=9) Global (n=17) Situational (n=11) Global (n=23) Situational (n=15) Event (n=5)
N sig. (%) N total
N sig.
(%)
N
total N sig. (%)
N
total N sig. (%)
N
total N sig. (%)
N
total
N sig.
(%) N total N sig. (%)
N
total
Amphetamine @ @ @ @ @ @ 0 (0%) 1 @ @ 0 (0%) 2 @ @
Cannabis 4 (57%) 7 1 (50%) 2 1 (20%) 5 1 (50%) 2 4 (25%) 12 4 (57%) 7 1 (25%) 4
Cocaine 1 (25%) 4 @ @ 0 (0%) 2 1 (100%) 1 4 (80%) 5 0 (0%) 3 0 (0%) 1
Crack cocaine @ @ @ @ @ @ @ @ 1 (50%) 2 @ @ @ @
Crystal methamphetamine 4 (57%) 7 @ @ 5 (83%) 6 @ @ 4 (67%) 6 2 (100%) 2 3 (100%) 3
Ecstasy 1 (25%) 4 @ @ 0 (0%) 1 1 (100%) 1 @ @ 1 (33%) 3 1 (25%) 4
Erectile Dysfunction Drugs
(EDDs) 2 (67%) 3 1 (100%) 1 2 (100%) 2 1 (100%) 1 3 (60%) 5 1 (50%) 2 3 (100%) 3
GHB/GBL 2 (50%) 4 @ @ 2 (67%) 3 1 (100%) 1 1 (100%) 1 @ @ 3 (100%) 3
Heroin 1 (100%) 1 @ @ @ @ @ @ @ @ 0 (0%) 1 @ @
Ketamine 2 (50%) 4 @ @ 0 (0%) 2 1 (100%) 1 @ @ @ @ 0 (0%) 1
Mephedrone 2 (100%) 2 @ @ 0 (0%) 1 @ @ @ @ @ @ @ @
Poppers 9 (69%) 13 1 (100%) 1 5 (56%) 9 2 (100%) 2 11 (92%) 12 2 (50%) 4 3 (100%) 3
Chemsex grouped 4 (67%) 6 6 (86%) 7 5 (100%) 5 9 (100%) 9 3 (100%) 3 8 (100%) 8 @ @
Note: Studies measured multiple drugs for the same outcomes, therefore column totals do not add up to the total number of studies
*One study included both global and situational associations so is counted in both sub@total
25
Figure 1. Flow diagram of the identification process.
26
<8D%<$<9%<ED2
> -">C0- #
""
O
'988D<D72
%-"" ""
O
9>#"#?"5- "
O
81<7<*<1<%I2
$--&.> -"?
O
<D91B'<ED2
"" -
9>#"#?#-?
O
$--&.> -". -:
"#"
O
P #-- ;#
P"Q P -(
-P#8(-";--&.
<8D%<$<9%<ED2
> -">C0" (;#
(0-
O
Figure 2. Map of countries with studies on sexualised drug use among men who have sex
with men (MSM), trans women, or both, and those that have reported chemsex drug use
included in the review.
27
