Men who have sex with men more often chose daily than event‐driven use of pre‐exposure prophylaxis: baseline analysis of a demonstration study in Amsterdam

Abstract Introduction The Amsterdam PrEP project is a prospective, open‐label demonstration study at a large sexually transmitted infection (STI) clinic. We examined the uptake of PrEP; the baseline characteristics of men who have sex with men (MSM) and transgender persons initiating PrEP; their choices of daily versus event‐driven PrEP and the determinants of these choices. Methods From August 2015 through May 2016, enrolment took place at the STI clinic of the Public Health Service of Amsterdam, the Netherlands. MSM or transgender persons were eligible if they had at least one risk factor for HIV infection within the preceding six months. Participants were offered a choice between daily or event‐driven use of tenofovir/emtricitabine. Baseline data were analysed using descriptive statistics and multivariable analysis was employed to determine variables associated with daily versus event‐driven PrEP. Results Online applications were submitted by 870 persons, of whom 587 were invited for a screening visit. Of them, 415 were screened for eligibility and 376 initiated PrEP. One quarter (103/376, 27%) chose event‐driven PrEP. Prevalence of bacterial STI was 19.0% and mean condomless anal sex (CAS) episodes in the preceding three months were 11. In multivariable analysis, older age (≥45 vs. ≤34, aOR 2.1, 95% CI 1.2 to 3.9), being involved in a steady relationship (aOR 1.7, 95% CI 1.0 to 2.7), no other daily medication use (aOR 0.6, 95% CI 0.3 to 0.9), and fewer episodes of CAS (per log increase aOR 0.7, 95% CI 0.6 to 0.9) were determinants for choosing event‐driven PrEP. Discussion PrEP programmes are becoming one of the more important intervention strategies with the goal of reducing incident HIV‐infection and we were unable to accommodate many of the persons applying for this study. Offering a choice of dosing regimen to PrEP users may enable further personalization of HIV prevention strategies and enhance up‐take, adherence and cost‐effectiveness. Conclusions The majority of participants preferred daily versus event‐driven use. Within this majority, a high number of CAS episodes before PrEP initiation was reported and we observed a high prevalence of STI. Determinants of choosing event‐driven PrEP were older age, fewer CAS episodes, no other daily medication use, and involved in a steady relationship.


| INTRODUCTION
In the Netherlands, men who have sex with men (MSM) accounted for 65% of new HIV diagnoses in 2015 and HIV incidence has not been markedly declining [1]. These findings indicate the urgent need for new methods of HIV prevention.
The use of daily oral tenofovir disoproxil fumarate combined with emtricitabine (TDF/FTC) as pre-exposure prophylaxis (PrEP) effectively protects MSM against HIV infection [2]. The Ipergay study showed comparable efficacy in MSM when using non-daily event-driven dosing of PrEP (i.e. taking PrEP before and after sexual contact) [3,4].While daily PrEP use for at-risk populations has been included in several guidelines (e.g. World Health Organization and U.S. Centers for Diseases Control and Prevention) [5,6], event-driven PrEP use has been recommended in French guidelines and the most recent European Aids Clinical Society guidelines [7,8].
Issues that impede PrEP implementation in high-income countries include uncertainty about the numbers and characteristics of PrEP users, as well as costs related to its use [9,10]. Event-driven use could improve cost-effectiveness since fewer pills are required [11,12]. For client-centred and cost-effective implementation, more information is needed on the uptake of both daily and event-driven dosing regimens in real-life settings and on the determinants of choosing between these regimens.
We started the Amsterdam PrEP (AMPrEP) demonstration project with the aim of assessing uptake of daily and event-driven PrEP, at the participant's discretion, among HIV-negative MSM and transgender persons at increased risk for HIV infection. This project was part of a comprehensive HIV-reduction package offered at a large STI clinic. We assessed interest in PrEP use by reporting the number of study applicants during the study application period and evaluated baseline characteristics including STI prevalence and preference for daily versus event-driven PrEP use. In addition, we examined the predominant self-reported motives for PrEP use and choice of regimen, and analysed determinants for choosing event-driven PrEP use.

| Application procedures and study site
The AMPrEP project, a prospective, longitudinal, open-label demonstration study, began in June 2015. A press release announced the start of the study and the number of available spots (n=370), followed by large-scale media attention at the local and national level. We reached out to transgender people through their community organizations and via the centre for prostitution and health, which offers information, advice and support to sex workers including transgender persons. However, we did not launch a media campaign to recruit participants. The main inclusion criteria were published online and people interested in participating applied using a web-based form. On this form, we asked them to select all applicable inclusion criteria. Since a high number of interested people was expected and we wanted to allow ample time for individuals to express their interest in participation, we decided not to include individuals on a first-come, first-serve basis, but instead opted for a four-week application period (initial application period). All applicants then received a randomlyassigned rank number for a subsequent screening visit. No stratified sampling of at-risk subgroups (e.g. based on sexual behaviour characteristics, for young MSM, or transgender persons) was employed. During the first screening visit and prior to any assessments, applicants provided written informed consent to participate in the study. Visits took place from 3 August 2015 to 30 May 2016 at the STI outpatient clinic of the Public Health Service of Amsterdam (GGD). This is the largest STI clinic in the Netherlands, performing about 45,000 consultations yearly; 38% of clients are MSM. In 2017, we diagnosed 92 HIV infections, 77 of these were MSM.
The study was approved by the ethics board of the Academic Medical Center, Amsterdam, the Netherlands (NL49504.018.14). The AMPrEP project is part of the HIV Transmission Elimination AMsterdam (H-TEAM) initiative, a multidisciplinary and integrative approach to stop the urban epidemic (Hteam.nl). The AMPrEP study was registered at the online Dutch trial registry (registration number NTR5411) and the protocol is available online [13].

| Eligibility criteria
HIV-negative MSM and transgender persons (male-to-female and female-to-male) who have sex with men were eligible if they were at least 18 years of age and had one or more of the following risk factors for HIV infection within six months prior to the screening visit: condomless anal sex (CAS) with casual partners, at least one bacterial STI (i.e. syphilis, rectal or urethral chlamydia or gonorrhoea), use of post-exposure prophylaxis (PEP) after a sexual risk incident, or an HIV-positive sexual partner with a detectable viral load. We excluded people if they tested positive for hepatitis B surface antigen (HBsAg); had creatinine clearance rate according to the Cockroft-Gault equation less than 60 mL/minute [14]; had more than trace protein in urine, or had other contra-indications for TDF/FTC use (see protocol [13]). In case of symptoms or signs suggestive of acute HIV infection, we performed rapid point-of-care HIV RNA testing.

| PrEP dosing regimen
Participants were proposed a choice of free-of-charge daily or event-driven PrEP (TDF/FTC) use. Also, no costs were incurred on the participants for renal function and STI testing. Daily use consisted of one pill every day and event-driven use two tablets taken between 24 hours and 2 hours before CAS, followed by one tablet every 24 hours up to 48 hours after the last episode of sexual intercourse [3]. In the decision-making process, we used shared and informed decision making, (i.e. we explained both PrEP regimens and answered questions), thus providing the necessary information for participants to select the option most-adapted to their lifestyle. No advice was given on the suitability of their decision of PrEP regimen. In addition, we informed participants that switching PrEP regimen was allowed at each subsequent study visit.

| Study visits
At the screening visit, eligibility criteria were verified. Urine, blood and pharyngeal/anal samples were taken to test for STI, hepatitis B virus (HBV), HIV, serum creatinine and urine protein ( Figure 1). One to four weeks later at PrEP initiation, we verified that all diagnosed STI had been treated according to clinic protocols, and if not we provided immediate treatment; eligibility criteria were confirmed, a self-administered questionnaire was completed, and blood was taken to test for HIV, hepatitis C virus (HCV) and HBV infections. Participants were asked to return for follow-up visits one month after the PrEP initiation visit and every three months thereafter. They will be provided with PrEP until June 2018.

| Diagnostic testing
We performed HIV and STI testing according to routine STI clinic protocols. This included HIV antigen and antibody testing (LIAISON XL Murex HIV Ag/Ab, Diasorin, Saluggia, Italy) with immunoblot confirmation (INNO_LIPA HIV I/II Score, Fujirebio, Ghent, Belgium), and serology testing for syphilis (LIAISON Treponema Screen, Diasorin). If indicated, direct microscopy (dark-field or Gram-stained smear) was performed for samples taken from rectal or urethral ulcers or discharge to detect the presence of Neisseria gonorrhoeae or Treponema pallidum. Nucleic amplification testing for Chlamydia trachomatis and Neisseria gonorrhoeae (Aptima combo 2, Hologic Gen-Probe Inc., San Diego, USA) was performed on urine, anal and pharyngeal swabs. Samples from participants without a history of HBV vaccination were screened for anti-Hepatitis B core and anti-Hepatitis B surface antibodies (LIAISON Anti-HBc and anti-HBs II, Diasorin) and, if negative, participants were offered HBV vaccination.
In addition to the routine STI clinic protocol, blood taken at the PrEP initiation visit was tested for anti-HCV antibodies (Architect anti-HCV, Abbott Laboratories, Wiesbaden, Germany) with immunoblot confirmation (INNO-LIPA HCV Score, Fujirebio, Ghent, Belgium) as was previously described [15].

| Participant characteristics and risk behaviour
At the PrEP initiation visit, participants completed a self-administered questionnaire on risk behaviour over the past three months, socio-demographic characteristics (e.g. postal code, sexual preference, education, employment, income and living situation), and their main reason for participating in the AMPrEP project (from a list of 10 predefined options). Questions were asked on concern of acquiring HIV using a Likert scale from one (very concerned) to seven (not concerned at all). Educational level was divided into low (primary school and lower secondary vocational education), medium (intermediate and higher secondary general education, senior secondary vocational and preuniversity education) or high (higher professional or university education). Regarding sexual behaviour, participants were asked about number and type (steady, known casual, or anonymous) of sexual partners, the number and type (insertive or receptive) of anal sex episodes, and condom use. Using lay terms, questions were asked on substance use, both in general and around the time of sex, and whether they had injected any drugs. Chemsex  5) At the PrEP initiation visit, we took samples for pooled HIV, hepatitis B and hepatitis C virus testing; we performed a semi-structured interview on participants' main reason to start PrEP and for choosing daily or event-based PrEP. We also collected self-reported questionnaire data.
was defined as using gamma-hydroxybutyrate (GHB)/gammabutyrolactone (GBL), mephedrone or crystallized methamphetamine around the time of sex [16]. Finally, participants were asked if they routinely used medication and pharmaceutical interactions with TDF/FTC were checked.

| Other measurements at PrEP initiation
Participants were asked the most important reason to participate in the AMPrEP project and to complete the Mental Health Inventory screening test (MHI-5), Alcohol Use Disorders Identification Test (AUDIT), Drug Use Disorders Identification Test (DUDIT), and sexual compulsivity scale (SCS) [17]. The following score cutoffs were used to identify specific problems: 60 for MHI-5 (i.e. lower than 60 suggests an anxiety or depressive mood disorder) [18,19], eight for AUDIT (i.e. eight and higher suggests harmful alcohol use) [20], eight for DUDIT (i.e. eight and higher suggests drug-related problems) [21] and 24 for SCS (i.e. 24 and higher suggests a greater impact of sexual thoughts on daily functioning and an inability to control sexual thoughts or behaviours) [22,23].
Semi-structured interviews were used to identify the two most important motives for choosing daily versus event-driven PrEP at the PrEP initiation visit.

| Statistical analyses
First, we described the number of applicants, number screened and number initiating PrEP. Second, we reported the socio-demographic characteristics, risk behaviour, STI diagnoses, choice of PrEP dosing regimen and motives for participation and choice. Third, we compared characteristics between participants who chose daily PrEP compared to event-driven PrEP using Chi-squared tests for categorical data and student t-test or rank sum tests for continuous data. Fourth, variables associated at p < 0.15 in univariable analysis were included in a multivariable logistic regression model while forcing inclusion of age as a covariable. In case of collinearity between variables (e.g. number of anal sex partners and number of anal sex episodes), only one of the collinear variables was included. Number of CAS episodes was natural log transformed. Using a backward step-wise procedure, we retained variables with a p < 0.05 (according to the likelihood ratio test) in the final multivariable model. All analyses were performed using STATA Intercooled 13.1 (STATA Corporation, College Station, TX, USA).

| Socio-demographic characteristics and STI
prevalence Two of the 376 participants self-identified as male-to-female transgender persons and 374 as MSM ( Table 1). Most of the participants were white (85.1%) and had a high educational level (76.1%). The median age was 39 years [interquartile range (IQR), 32 to 48].
At PrEP initiation, 71 of 376 (19.0%) were diagnosed with a bacterial STI. Overall, 42 (11.2%) cases of anal gonorrhoea or chlamydia and 6 (1.6%) cases of syphilis were diagnosed. Moreover, 18 (4.8%) were positive for anti-HCV (n=17) and/or HCV RNA (n=15). Of 228 participants who had tested at the STI clinic within the preceding year, 103 (45.2%) had been diagnosed with a bacterial STI. Two people initiating daily PrEP tested HBsAg-positive, but were enrolled despite this exclusion criterion. One had been using event-driven PrEP, unmonitored and obtained elsewhere, while the second was mistakenly included. We referred both to specialist care to have their HBV infection monitored. Furthermore, we allowed continuation of study participation and advised to use PrEP daily to prevent hepatic flares when PrEP is interrupted.

| Sexual risk behaviour, drug use and other measurements
Characteristics of risk behaviour and drug use in the three months prior to PrEP initiation are described in Table 1. The median number of anal sex partners in the three months before PrEP initiation was 15 [IQR 6 to 30] and the median number of CAS episodes was 11 (IQR 4 to 23). Of all participants, 282 (75.0%) reported receptive CAS, and 203 (54.0%) reported at least ten CAS episodes in the preceding three months.
Of all participants, 156 (42.0%) reported chemsex and 15 (4.0%) reported injecting drug use in the preceding three months. More than one third (35.7%) scored eight or higher on DUDIT and 28.0% scored eight or higher on AUDIT, suggesting drug and alcohol use disorder respectively. According to the MHI-5, 20.7% scored below 60, suggesting a depressive mood and/or anxiety disorder.   contracting HIV (n=98, 26.1%) (Figure 2A). The two main reasons for PrEP regimen choice were daily PrEP -"it is easier to use" and "it is difficult to predict when I will have sex;" eventdriven PrEP -"I am able to predict sex" and "I am not often at risk" (Figure 2B and C).

| Determinants for choosing event-driven PrEP use
In univariable analyses, older age was positively associated with choosing event-driven PrEP, whereas persons with higher numbers of anal sex partners and CAS episodes and more eligibility criteria were less likely to choose event-driven PrEP (Tables 1 and 2). Several other factors, including educational level and a higher score on the screening tool for drug disorder were not associated with choice of PrEP regimen. In multivariable analyses, older age (≥45 vs.≤34, adjusted odds ratio (aOR) 2.1, 95% CI 1.2 to 3.9) and being involved in a steady relationship (aOR 1.7, 95% CI 1.0 to 2.7) were positively associated with event-driven PrEP; daily medication use (aOR 0.6, 95% CI 0.3 to 0.9) and more CAS episodes in the preceding three months (per log increase  aOR 0.7, 95% CI 0.6 to 0.9) were negatively associated with event-driven PrEP.

| DISCUSSION
In the Amsterdam PrEP project, the demand for PrEP among MSM largely exceeded capacity. The majority of the participants in this Dutch demonstration project preferred daily versus event-driven use. Within this majority, a high number of CAS episodes before PrEP initiation was reported and a high prevalence of STI, including HCV was observed. Furthermore, 42% reported engaging in chemsex. Determinants of choosing event-driven PrEP were older age, fewer CAS episodes in the preceding three months, no other daily medication use, and being in a steady relationship.
To the best of our knowledge, we are the first to report on the preferences between daily and event-driven PrEP among MSM and the reasons for such preferences. Although most people opted for daily use, about one quarter preferred event-driven use. People who opted for event-driven PrEP reported fewer CAS episodes and thus would require PrEP less often for protection; they were less likely to take daily medication whereby daily PrEP could be conveniently integrated. Providing individuals a choice between daily or eventdriven PrEP would appear to enable further personalization of their HIV prevention strategy and could accompany increased adherence and resulting protection. In addition, as event-driven users take fewer PrEP pills, offering this choice has the potential to reduce cost at the individual and population level alike, and improve cost-effectiveness [11,12]. In a PrEP rollout study in France, uptake of daily PrEP is lower than in our project: 41% of individuals preferred daily PrEP and 59% eventdriven PrEP [4]. The difference in regimen preference between the French study and our own might reflect the fact that French providers and MSM were more familiar with event-driven dosing from the randomized-controlled trial Ipergay in France [3].
The demand for PrEP was reflected by the large number of submitted applications to participate in AMPrEP over a short time period. Nevertheless, we clearly stated to participants that late applicants were unlikely to be enrolled due to limited capacity. Potential applicants interested in PrEP could have perceived their chance of inclusion as low and thus were discouraged from participating, implying that the number of those actually interested in initiating PrEP in Amsterdam are higher than the number of applicants to the AMPrEP project. Currently, few data on intention to use PrEP in the Netherlands have been published [24,25] and only crude estimates of the number of MSM willing to take PrEP are available [26,27]. The participants initiating PrEP in this study were indeed at substantial risk for HIV infection, as shown by a high median number of CAS episodes and high STI prevalence. The proportion with a bacterial STI (19.0%) at PrEP initiation corresponded to other PrEP cohorts (15.9% to 28.8%) [3,28]. However, the prevalence of HCV infection was higher in our study [3,15,29], indicating that STI testing strategies should include HCV among HIV-negative MSM initiating PrEP.
The proportion of men reporting chemsex in the three months before PrEP initiation is high and in line with two other European PrEP cohorts [3,19] and with studies in the larger MSM community [30,31]. Moreover, in our study, a large proportion scored low on MHI-5 or high on the DUDIT, AUDIT, or SCS screening tools, indicating risk for anxiety or depressive mood disorder, alcohol and drug use disorder, or  [3,19,33,34]. To increase their inclusion, future studies will require additional and specifically-tailored PrEP campaigns and programmes based on research identifying the specific needs of transgender people at risk for HIV infection (e.g. providing PrEP at gender-affirming care facilities) [35,36].
One limitation of our study is that, despite the project being conducted in an STI clinic offering routine sexual healthcare, choices for daily and event-driven PrEP may not reflect those in the population at large. It should be mentioned that PrEP was offered free-of-charge, which is currently not the case in most countries. More people might opt for event-driven PrEP under the premise that future practice would require payment. Furthermore, our project is the first to offer access to a free-of-charge PrEP programme in the Netherlands and as a result, we may have attracted a higher number of early adopters who may not be representative of the wider MSM community.
The major strength of this study is the knowledge generated on PrEP use and choice of dosing regimens in a realworld setting. So far, little data have been available, especially in Europe. The STI clinic in Amsterdam represents an ideal location for conducting this PrEP project given the frequent previous STI testing among participants (Table 1). Our initially high number of online applications provides further support for implementation of PrEP at STI clinics.
In conclusion, we were unable to accommodate many of the persons applying for this study in the Netherlands. Participants were willing to start daily or event-driven PrEP as part of their routine STI care. They were able to choose a dosing regimen and the larger preference towards daily use was probably related to their risk behaviour. Offering a choice of dosing regimen may enhance uptake, adherence, and cost-effectiveness. The results of this study are important for policy makers, health insurers, and healthcare professionals concerned with implementation of efficient and cost-effective PrEP programmes in high-income countries. Until the availability of new interventions, such as an HIV vaccine, PrEP programmes are becoming one of the more important intervention strategies with the goal of reducing incident HIV infection. Therefore, it is of importance to provide the option between daily and event-driven PrEP for all people at high risk of acquiring HIV.

C O M P E T I N G I N T E R E S T S
We received the study drug for the Amsterdam PrEP study from Gilead Sciences. EH: my institute received financial reimbursement for time I spent serving on advisory boards of Gilead Sciences, and a speaker fee from Janssen-Cilag. The authors thank the following persons for their invaluable support to this study: Ineke Stolte, Gerben Rienk Visser, Linda May, Thijs van de Laar, Paul Oostvogel, Sylvia Bruisten, Anders Boyd, Adriaan Tempert, Kees de Jong, Ilya Peters, Myra van Leeuwen, Princella Felipa, Lucy Phillips, as well as the teams of the STI outpatient clinic and research department of GGD Amsterdam. Furthermore, we thank the members of the advisory board and the community engagement group of AMPrEP, and all of those who contributed to the H-team (see Annex S1). Finally, we would like to acknowledge Dr. Han Fennema, the former medical director of the GGD Amsterdam, who died in 2016. Dr. Fennema initiated the PrEP research activities at our institute and was always a strong supporter of PrEP and the team.