Secondary Logo

Journal Logo

Prevention Research

Evaluation of Drug Abuse by Hair Analysis and Self-Reported Use Among MSM Under PrEP: Results From a French Substudy of the ANRS-IPERGAY Trial

Chas, Julie MDa; Bauer, Rebecca MScb; Larabi, Islam Amine PharmDc; Peytavin, Gilles PharmD, PhDd; Roux, Perrine PhDe,f; Cua, Eric MDg; Cotte, Laurent MDh; Pasquet, Armelle MDi; Capitant, Catherine MDb; Meyer, Laurence MD, PhDb,j; Raffi, Francois MD, PhDk; Spire, Bruno MD, PhDe,f; Pialoux, Gilles MD, PhDa,l; Molina, Jean-Michel MD, PhDm,n,o; Alvarez, Jean-Claude PharmD, PhDc

Author Information
JAIDS Journal of Acquired Immune Deficiency Syndromes: April 15, 2021 - Volume 86 - Issue 5 - p 552-561
doi: 10.1097/QAI.0000000000002610
  • Free



Drug use is on the rise among men who have sex with men (MSM), particularly in sexual contexts, with the emergence of a new class of drugs, in addition to conventional drugs, called new psychoactive substances (NPS). NPS are natural and synthetic recreational molecules, which mimic the effects of traditional drugs of abuse (eg, cocaine, amphetamine, and ecstasy). Their molecular structures are similar, but not completely identical. This specificity allows them, at least in the short term, to circumvent drug legislation; some are classified others have no clear legal status. Generally purchased on the Internet, NPS are known either by their chemical names or through trade names. Nearly 100 NPS have appeared in recent years.1,2 Both the scientific community and users often lack adequate information on the effects of these new drugs when used alone or in combination. Synthetic cathinones are of particular concern because some drugs in this class are used instead of ecstasy without consumers' knowledge as their effects are close to those of ecstasy or other stimulants.3,4 Cathinones exposures have been associated with many side effects, both physical and psychological.5–9 Users may not always know what drug they are really taking, such as ecstasy that may contain synthetic cathinones.10 This drug use is also associated with risky sexual behavior in the context of chemsex,11–14 defined as the use of psychoactive drugs during sexual intercourse.15,16 This practice is called “slam” when drugs are being injected. To respond individually and collectively to worrying NPS consumption, it is important to precisely identify the drugs used and to offer users the most targeted multidisciplinary harm reduction counseling and care possible. Most studies evaluating recreational drug use among MSM are qualitative studies and measure drug consumption by self-questionnaires with many subjective associated biases.14,17 We used the Agence nationale de Recherches sur le sida et les hépatites virales (ANRS)-IPERGAY trial to test a new method of measuring drugs consumed by hair analysis that allows for objective, more reliable and accurate qualitative and quantitative measurements of drug use. Hair analysis has become a new routine tool for the detection of xenobiotic intake, including NPS.18,19 Many studies investigate the association of hair ART concentrations with virologic outcomes to determine thresholds of hair concentrations of various ARTs associated with virologic suppression.20 Hair analysis may provide a noninvasive, cost-effective assessment of drug abuse. The main advantage of this matrix over traditional ones such as urine or blood is its window of drug detection that is dramatically extended to weeks or months or even years according to the length of the strand (1 cm of hair corresponding approximately to 1 month of exposure). In parallel, drug use was also measured by self-questionnaires. The objective of our study was to qualitatively and quantitatively measure drug use among MSM under preexposure prophylaxis (PrEP) using 2 different measurement methods, self-report and direct measures, to better understand uses and to be able to respond collectively to risky practices with appropriate messages.


Study Population

This study was a substudy of the ANRS-IPERGAY trial, which showed efficacy of on-demand PrEP with tenofovir disoproxil fumarate plus emtricitabine in reducing HIV incidence among high-risk MSM.21–23 Following this result, an open-label study extension was implemented from November 4, 2014, to June 30, 2016, to provide all participants with tenofovir disoproxil fumarate plus emtricitabine until PrEP approval (full approval in France in January 2016). Briefly, eligible participants for inclusion in the ANRS-IPERGAY trial were HIV-negative men or women having sex with men, aged 18 years or older, and at high risk for HIV acquisition (defined as having condomless anal sex with at least 2 different partners during the past 6 months). Event-based dosing involved taking 2 tablets between 2 and 24 hours before potential exposure, a third tablet 24 hours later, and a fourth 24 hours after that, with daily dosing until at least 2 days after the last potential exposure as previously reported.21 During the ANRS-IPERGAY trial, participants completed online questionnaires at baseline, then every 2 months, and annually. Participants were also asked to provide hair samples every 4 months initially for antiretroviral monitoring. In November 2015, the ANRS-IPERGAY protocol was again amended to implement this substudy of qualitative and quantitative evaluation of the prevalence of conventional drugs and NPS consumption among MSM in French sites alone. Public health authorities and ethics committees in France (Committee for the Protection of Persons, Paris Saint-Louis) approved this amendment. The present substudy included all participants who provided written informed consent.

Substudy Period

The analysis period of the substudy was defined between the first hair sample collected on October 8, 2013, and the last hair sample collected on May 30, 2016.

Hair Analysis

Hair analyses were performed using at least 20 mg of sample. Each hair strand was decontaminated successively by dichloromethane and hot water. One hair strand can give several samples if the hair is long. Thus, when possible (hair length ≥3 cm and weight ≥20 mg) a segmental analysis (2-cm segments) was performed, 1 cm corresponding to 1 month of previous drug use. Hair of each segment was cut into small pieces of less than 2 mm. Samples were then extracted by liquid–liquid extraction before being analyzed by liquid chromatography coupled with high-resolution chromatography (LC-HRMS, Q-Exactive; ThermoFisher) and liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS, TSQ Vantage; ThermoFisher). LC-HRMS untargeted screening was realized by data-dependent analysis; compounds were identified thanks to a shared MS2 spectra library.24 Confirmation and quantification were performed using a LC-MS/MS-accredited method (Comité Français d'Accréditation, COFRAC) in multiple reaction monitoring mode. This method previously described25,26 was validated according to the European Medicines Agency's guidelines and allows to quantify the conventional drugs of abuse (opiates, cocaine and metabolites, amphetamines, and phenethylamines), 86 NPS, and therapeutic substances. Segmental hair analysis performed showed different patterns of consumption, ranging from rare intake (concentration <200 pg/mg) to regular intake (>1000 pg/mg). Hair samples of the participants were analyzed every 4 months to detect drugs (except acide GammaHydroxyButyrique/gamma-butyrolactone (GHB/GBL) and THC, because both required another hair sample that was not included in the original design of the ANRS-IPERGAY study, and poppers for technical reasons).


Participants completed online questionnaires at baseline, every 2 months visits, and annual visits. Baseline questionnaires collected data on sociodemographic and socioeconomic characteristics (age, educational level, and employment status), data on drug abuse, and data on sexual behavior during the previous 12 months. The every 2-month follow-up questionnaires collected data on drug abuse and sexual behavior during their most recent sexual intercourse. Annual questionnaires collected data on drug abuse and sexual behavior during the previous 12 months and during their most recent sexual intercourse. In July 2015, participants were asked to answer an additional section in the every 2-month follow-up questionnaire and annual questionnaire, which addressed the use of an extended list of psychoactive substances during their most recent sexual intercourse and during the previous 2 months (cannabis, ecstasy, cocaine, poppers, alcohol, GHB/GBL, ketamine/Special K, Viagra, and, as of July 2015, crack, heroin, methamphetamine/speed/crystal meth, diéthyllysergamide, or mephedrone/cathinone/méthylènedioxypyrovaléron/commercial name of cathinones/4-methyl-N-ethylcathinone). Quantitative assessment of drugs used was based on different patterns of use, ranging from rare intake (less than once a week) to regular intake (every day). For data analysis, all questionnaires completed during the substudy period (October 8, 2013, to May 30, 2016) were taken into account.

Drug Consumption Definition

For hair samples analysis, consumers were defined as participants with at least 1 hair sample containing at least 1 NPS or conventional drug (excluding therapeutics drugs, GHB/GBL, poppers, and THC). For questionnaire analysis, consumers were defined as participants with at least 1 reported drug use (ecstasy, cocaine, ketamine/Special K, GHB/GBL, and, as of July 2015, crack, heroin, methamphetamine/speed/crystal meth, diéthyllysergamide, or mephedrone/cathinone/méthylènedioxypyrovaléron/commercial name of cathinones/4-methyl-N-ethylcathinone) among all questionnaires collected during the analysis period of the substudy. Chemsex was defined as reporting to be under the effect of at least one of these substances during their most recent sexual intercourse. Recreational drug use was defined by drug use in general, whether in a festive or chemsex context.

Sexual Behavior

Data collected on sexual behaviors included type of partner (main partner, causal partner, or sex party); number of partners during in the past 2 months; number of sexual acts in the past 4 weeks; type of sexual practice (oral sex, insertive anal sex, or receptive anal sex); high-risk HIV exposure (condomless anal sex or not); and hardcore sexual practices in the last sexual act.

Sexually Transmitted Infection

Participants were tested every 6 months. Syphilis was diagnosed by a positive serological assay with treponemal and nontreponemal tests. Chlamydial and gonococcal infections were defined by a single positive polymerase chain reaction test or culture from at least 1 site (throat, urine, or anus). In case of a positive polymerase chain reaction result at multiple sites simultaneously with the same organism, a single infection was counted. Data collected on sexually transmitted infections (STIs) used case Report Form informatic (e-CRF) medical data.

Psychiatric Disorders

Data collected on psychiatric disorders used e-CRF medical data and were coded with Medical Dictionary for Regulatory Activities Terminology. The system organ classes associated with these events is psychiatric disorders, with 4 levels of severity: mild, moderate, severe, and life threatening.

Statistical Analysis

We considered all self-questionnaires between baseline and the date of the last hair sample. The percentages of responses to a question were calculated based on the number of participants who could answer the question. Comparisons were performed using the χ2 test or Fisher exact test for categorical variables, the Wilcoxon test for continuous variables, and the McNemar test for paired nominal data. The number of participants with at least 1 STI was calculated from the medical data of the e-CRF. The same applied to psychiatric events. All analyses were performed with software R version 3.5.2.



From February 22, 2012, to June 30, 2016, 429 participants were included in the ANRS-IPERGAY trial in 6 French centers and 1 Canadian center. The substudy was performed only in French centers for logistical reasons, thus including 381 participants. At the start of the substudy on january 4, 2016, of the remaining 210 participants screened for substudy, 44 were no longer followed in the main study, 41 refused to participate (reasons for refusal were not provided), 1 was infected with HIV, 18 had missing and 106 provided written informed consent. Of these 106 participants, only 69 had the sufficient hair matrix to be analyzed (Fig. 1).

Study profile. *For logistical and financial reasons. †Too short hair to be collected and analyzed. ‡Reasons for refusal not provided. §No sufficient hair matrix to be analyzed.

Characteristics of the Substudy Sample Versus Study Sample at Baseline

A total of 5898 questionnaires corresponding to 429 participants included in the ANRS-IPERGAY trial were available for analysis. No significant difference was found between individuals in the drug analysis substudy and other participants in the IPERGAY study, except for the circumcision that tended to be less frequent in individuals in the substudy (P = 0.045) (Table 1).

TABLE 1. - Baseline Characteristics (ANRS-IPERGAY Study, n = 429 Participants, 5898 Questionnaires)
Drug Substudy Population (n = 69 Participants) IPERGAY Not in Drug Substudy Population (n = 360 Participants) P
Age, yrs 34.7 (28–40.9) 35.2 (29.2–42.8) 0.208
Sexual orientation
 MSM 67 (97) 347 (96) 1.000
 Bisexual 2 (3) 13 (4)
Education level
 Postsecondary* 45 (65) 262/355 (74) 0.140
Unemployed 15 (22) 44/355 (12) 0.056
Use of recreational drugs 29 (42) 163/355 (44) 0.598
 In the context of chemsex practice 7/69 (10) 46/355 (13) 0.690
Sexual risk factors at baseline
 No. of partners in the past 2 mo§ 9 (5–15) 8 (5–17) 0.921
 No. of sexual acts in the past 4 wks§ 10 (5–16) 10 (5–17) 0.792
 Fisting in the last sexual act 2/51 (4) 28/220 (13) 0.084
 Hardcore sexual practices in the last sexual act 4/51 (8) 23/221 (10) 0.796
 Condomless anal sex in the last sexual act 18/41 (44) 55/185 (30) 0.116
Circumcision 8 (12) 80 (22) 0.045
At least 1 bacterial sexually transmitted disease in the past 12 mo# 23 (33) 92 (26) 0.182
Data are n (%), median (interquartile range), or n/N (%).
*Postbaccalaureate certificate achieved at the end of high school.
Recreational drugs used in general in the past 12 months or chemsex practice at most recent sexual encounter, including ecstasy, crack, cocaine, crystal (methamphetamine), speed (amphetamine), and GHB/GBL.
‡Missing answers to questions about the most recent sexual encounter.
Including only participants with questionnaires with a positive answer concerning partners in the past 2 months or sexual acts in the past 4 weeks.
Including only participants who had anal intercourse reported.
#Included syphilis serological testing by rapid plasma reagin confirmed with the use of a treponema-specific assay; Neisseria gonorrhoeae and Chlamydiae trachomatis were detected by polymerase chain reaction with urine samples and throat and anal swabs.

Prevalence of Drug Use Assessed by Hair Analysis

Using hair analysis, NPS and/or conventional drugs were detected among 53 of the 69 (77%) participants. A total of 219 segments corresponding to 137 hair samples were analyzed with a median hair sample per participant of 2 [1;3] among consumers and 1.5 [1;2] among nonconsumers, without a statistical difference. Thirty-two different molecules were detected in 3 categories: 15 NPS, 11 therapeutic agents, and 6 conventional drugs. The most commonly used drugs were cocaine in 47 of the 69 (68%) participants and 3,4-methylenedioxymethamphetamine (MDMA)/ecstasy in 31 of the 69 (45%) participants. NPS consumption (especially cathinones) was detected in 27 of the 69 (39%) participants. In these users, NPS use was associated with cocaine use in 25 of the 27 (93%) participants, MDMA in 20 of the 27 (74%) participants, or ketamine in 19 of the 27 (70%) participants, reflecting polyconsumption. The use of NPS and/or conventional drugs seems to be rare (hair concentration <200 pg/mg). Drugs used on a regular basis (hair concentration >1000 pg/mg) were more amphetamines and methamphetamines. Therapeutic agents were detected among 50 of the 69 (72%) participants. The most commonly used therapeutic agents were erectile drugs in 25 of the 69 (36%) participants, followed by ephedrine in 18 of the 69 (26%) participants, and among analgesics drugs, tramadol and codeine in 15 of the 69 (22%) and 9 of the 69 (13%) participants, respectively (Table 2).

TABLE 2. - Drugs Detected and Pattern of Consumption by the Hair Method (n = 69 Participants, 137 Hair Samples, 219 Hair Segments, 2 Classifications: Drugs Common to Questionnaire Method and Other Drugs Only Detected by the Hair Method)
Drugs Detected Participants With Drug Detected (/69*) Hair Samples With Drug Detected (/137) Hair Segments With Drug Detected (/219) Detected But Undefined§ Rare Intake (<200 pg/mg) Occasional Intake (200–1000 pg/mg) Regular Intake (>1000 pg/mg)
(/N Hair Segments With Drug Detected)
Drugs common to the questionnaires method
 Cocaine 47 (68) 82 (60) 137 (63) 75/137 (55) 29/137(21) 33/137 (24)
31 (45) 56 (41) 91 (42) 50/91 (55) 21/91 (23) 20/91 (22)
 Ketamine 26 (38) 41 (30) 69 (32) 58/69 (84) 6/69 (9) 5/69 (7)
8 (12) 13 (9) 20 (9) 5/20 (25) 8/20 (40) 7/20 (35)
22 (32) 28 (20) 42 (19) 29/42 (69) 8/42 (19) 5/42 (12)
 Erectile drugs
25 (36) 59 (43) 99 (45) 47/99 (47) 36/99 (36) 16/99 (16)
 Methoxetamine 4 (6) 7 (5) 10 (5) 4/10 (40) 5/10 (50) 1/10 (10)
 Methylphenidate 1 (1) 1 (1) 1 (1) 1/1 (100)
 Ethylphenidate 5 (7) 5 (4) 10 (5) 9/10 (90) 1/10 (10)
 Methiopropamine 3 (4) 5 (4) 8 (4) 6/8 (74) 1/8 (13) 1/8 (13)
 PMMA 3 (4) 7 (5) 12 (6) 10/12 (83) 2/12 (17)
 5F-PB22 1 (1) 1 (1) 2 (1) 2/2 (100)
 Phendimetrazine 1 (1) 1 (1) 1 (1) 1/1 (100)
 Phentermine 1(1) 1(1) 1 (1) 1/1 (100)
 N-methyl-2AI 1 (1) 1 (1) 2 (1) 2/2 (100)
 Analgesic drugs
  Codeine 9 (13) 16 (12) 26 (12) 24/26 (92) 2/26 (8)
  Nefopam 7 (10) 11 (8) 17 (8) 16/17 (94) 1/17 (6)
  Dextropropoxyphene 4 (6) 5 (4) 6 (3) 6/6 (100)
  Pholcodine 4 (6) 5 (4) 7 (3) 4/7 (57) 1/7 (14) 2/7 (29)
  Nalmefene 1 (1) 1 (1) 1 (1) 1/1 (100)
  Tramadol 15 (22) 27 (20) 42 (19) 20/42 (48) 15/42 (36) 7/42 (17)
  Butorphanol 1 (1) 1 (1) 1 (1) 1/1 (100)
  Ephedrine 18 (26) 23 (17) 33 (15) 28/33 (85) 5/33 (15)
Total conventional drug and/or NPS detected 53 (77) 110 (80) 184 (84) 15/184 (8) 89/184 (4) 38/184 (21) 42/184 (23)
Total therapeutics drugs detected 50 (72) 108 (79) 179 (82) 6/179 (3) 60/179 (34) 62/179 (35) 51/179 (28)
Data are n (%).
Rows with bold values are conventional drugs.
Rows with italicized values are NPS (*cathinones).
Rows with bold and italicized values are therapeutics drugs.
*Of the69 participants.
Of the 137 hair sample.
Of the 219 hair segments.
§Not quantified (standard not available at the time of the detection).
Amphetamine manufacturing component.
MDPV, méthylènedioxypyrovaléron.

Prevalence of Drug Use Assed by Questionnaire Analysis

Using questionnaire analysis, NPS and/or conventional drugs were reported in 39 of the 69 (57%) participants. A total of 1061 questionnaires were analyzed. Ten different drug classes were detected in 3 categories: 8 conventional drugs, 1 NPS, and 1 therapeutic drug. The most commonly used drugs are conventional drugs with cocaine in 31 of the 69 (45%) participants, MDMA/ecstasy in 29 of the 69 (42%) participants, and GHB/GBL in 25 of the 69 (36%) participants. NPS consumption was reported in 16 of the 69 (23%) participants. The use of NPS and/or conventional drug was mostly rare (less than once a week). Erectile drugs were reported in 34 of the 69 (49%) participants (Table 3).

TABLE 3. - Drugs Reported and Pattern of Consumption by the Questionnaire Method (n = 69 Participants, 1061 Questionnaires, 2 Classifications: Drugs Common to the Hair Method and Others Drugs Only Detected by the Questionnaire Method)
Drugs Detected Participants With Drug Reported (/69*) Questionnaires With Drug Reported (/1061) Rare Intake (<1/wk) Occasional Intake (>1/wk) Regular Intake (Daily)
(/N Questionnaires With Drug Reported)
Drugs common to the hair method
 Cocaine 31 (45) 129 (12) 94/129 (73) 27/129 (21) 8/129 (6)
29 (42) 102 (10) 90/102 (88) 8/102 (8) 4/102 (4)
 Special K
15 (22) 33 (3) 31/33 (94) 1/33 (3) 1/33 (3)
16 (23) 35 (3) 28/35 (3) 6/35 (17) 1/35 (3)
16 (23) 58 (5) 40/58 (69) 17/58 (29) 1/58 (2)
 Total conventional drug and/or NPS reported 39 (57) 642 (61) 121/642 (19) 34/642 (5) 9/642 (1)
 Sildenafil (erectile drug) 34 (49) 78 (7) 50/78 (64) 17/78 (22) 11/78 (14)
 GHB/GBL 25 (36) 77 (7) 67/77 (87) 9/77 (12) 1/77 (1)
 LSD 6 (9) 9 (1) 9/9 (100) 0 0
 Crack 0 0 0 0 0
 Heroin 0 0 0 0 0
Data are n (%).
Rows with bold values are conventional drugs.
Rows with italicized values are NPS (*cathinones).
Rows with bold and italicized values are therapeutics drugs.
*Of the 69 participants.
Of the 1061 questionnaires.
LSD, diéthyllysergamide; MDPV, méthylènedioxypyrovaléron; NRJ3, commercial name of cathinones.

Similar data between hair and questionnaire were observed for 49 of the 69 (71%) participants. According to the McNemar test between drug use data from the hair analysis compared with questionnaire analysis, hair analysis detected more conventional drugs and/or NPS use (P = 0.004), but there was no difference between the 2 methods for detecting erectile drugs (P = 0.070).

Characteristics of the Study Sample (Consumers Versus Nonconsumers)

A total of 1061 questionnaires corresponding to 69 participants included in our substudy population were available for analysis: 848 questionnaires among 53 consumers and 213 questionnaires among 16 nonconsumers. At baseline, the median (interquartile range) age of the 53 consumers was 34.5 years (28.3–40.9); 35 (66%) had completed higher education and 44 (83%) were working. Fifty-one participants (96%) were MSM and 2 (4%) were bisexual; 7 (13%) were circumcised. For the follow-up, results were derived from all follow-up questionnaires considered during the period of the substudy. Drug use was significantly associated with a higher number of sexual partners in the past 2 months (P ≤ 0.001), more often casual partners (P ≤ 0.001), condomless anal sex (P ≤ 0.005), hardcore sexual practices (fisting and sadomasochistic practices) (P ≤ 0.001), and a higher number of STIs with at least 1 syphilis-type STI, gonococcal anal/genital/pharyngeal STI, or chlamydia anal/genital/pharyngeal STI (P ≤ 0.02). There was 1 hepatitis C virus infection among consumers and no HIV infection in either group. Practice of chemsex was significantly higher among consumers versus nonconsumers (P ≤ 0.05). Thirty-two psychiatric disorders were reported in 19 consumers: 8 with 13 moderate disorders (4 depression, 4 sleep disorder, 3 anxiety, and 2 nightmare), 13 with 16 severe disorders (8 depression, 2 sleep disorder, 1 affective disorder, 1 alcoholism, 1 anxiety, 1 burnout syndrome, 1 loss of libido, and 1 suicidal ideation), and 3 with 3 life threatening disorders (2 suicide attempt and 1 alcoholism). Three psychiatric disorders were reported in 4 nonconsumers: 3 with 3 moderate disorders (1 depression, 1 nightmare, and 1 emotional disorder) and 1 with 1 severe disorder (1 anxiety). The rate of psychiatric events in consumers was higher than in nonconsumers without any significant difference being established (Table 4).

TABLE 4. - Sociodemographic, Economic, Drugs Use, Sexual Behaviors and Psychiatric Characteristics at Baseline and During Follow-up by the Hair Method (ANRS-IPERGAY Substudy, n = 69 Participants, 1061 Questionnaires)
Consumers (n = 53 Participants) Nonconsumers (n = 16 Participants) P
At baseline
 Age, yrs 34.5 (28.3–40.9) 35.4 (27.6–40.4) 0.915
 Sexual orientation
  MSM 51 (96) 16 (100) 1.000
  Bisexual 2 (4) 0 (0)
 Circumcision 7 (13) 1 (6) 0.671
 Education level (postsecondary*) 35 (66) 10 (62) 0.774
 Unemployed 9 (17) 6 (15) 0.096
 Use of recreational drugs including GHB/GBL in the past 12 mo 26 (49) 3 (19) 0.043
 Use of recreational drugs excluding GHB/GBL in the past 12 mo 26 (49) 2 (12) 0.010
 In the context of chemsex practice including GHB/GBL in the past 12 mo 6 (11) 1 (6) 1.000
During follow-up
 At least 1 use of recreational drugs including GHB/GBL 37 (70) 4 (25) 0.003
 At least 1 use of recreational drugs excluding GHB/GBL 36 (68) 3 (19) 0.001
 At least 1 in the context of chemsex practice including GHB/GBL 22 (42) 2 (13) 0.039
 At least 1 bacterial sexually transmitted disease§ 44 (83) 7 (44) 0.003
  At least 1 Neisseria gonorrhoeae 26 (49) 5 (31) 0.259
  At least 1 Chlamydia trachomatis 30 (57) 2 (13) 0.003
  At least 1 syphilis 12 (23) 1 (6) 0.272
 Psychiatric disorder 19 (36) 3 (19) 0.237
  Moderate 8 (15) 3 (19)
  Severe 13 (25) 1 (6)
  Life threatening 3 (6) 0
 No. of partners in the past 2 mo 8 (4–15) 4 (2–9) <0.001
 No. of sexual acts in the past 4 weeks 10 (5–15) 7 (3–11) 0.015
 Sexual behaviors at most recent sexual encounter N = 701 questionnaires N = 183 questionnaires
  Type of partner
   Main partner 172 (25) 84 (46)
   Casual partner 468 (67) 91 (50) <0.001
   Sex party 61 (9) 8 (4)
  Sexual practices#
   Oral sex 110 (16) 25 (14)
   Insertive anal sex 251 (36) 57 (31) 0.336
   Receptive anal sex 339 (48) 99 (54)
  High-risk HIV exposure: condomless anal sex**
   No 413 (59) 129 (71)
   Yes 177 (25) 27 (15) 0.005
   No anal sex 111 (16) 26 (14)
  Hardcore sexual practices††
   No 582 (83) 171 (93)
   Yes 116 (17) 11 (6) <0.001
Data are n (%), median (interquartile range), or n/N (%).
*Postbaccalaureate certificate achieved at the end of high school.
Recreational drugs used reported by questionnaire including ecstasy, crack, cocaine, crystal (methamphetamine), speed (amphetamine), and +/− GHB/GBL.
‡One missing answers to questions about most recent sexual encounter.
All types of sexual acts are included (insertive or receptive anal sex and oral sex).
Only follow-up questionnaires with a positive answer concerning partners in the past 2 months or sexual acts in the past 4 weeks.
#Three missing values.
**One missing value.
††Sadomasochistic practices, fisting, and 21 missing values.


This study describes qualitatively and quantitatively drug use, licit or illicit, by MSM, volunteers under PrEP enrolled in the ANRS-IPERGAY study, through hair analysis and questionnaires. In PrEP studies published so far, drug use among MSM was only assessed using self-administered questionnaires.14,27–30

The overall prevalence of drug use detected by the hair method (53/69, 77%) in our study was higher than reported by the questionnaires in the various PrEP trials among high-risk MSM (43%–77%),21,22,27,28 in surveys conducted among MSM outside PrEP studies (18%–60.4%)8,14,31–33 and in our own ANRS-IPERGAY substudy (39/69, 57%). The difference in the prevalence rates observed between the drugs detected in hair and those reported in the self-questionnaires can be explained by different reasons: (1) Until July 2015, the bimonthly ANRS-IPERGAY questionnaires surveyed only drug use at the last sexual intercourse, which may explain some negative responses from participants about this 1-time use, but does not eliminate possible use at a time other than the last sexual intercourse. (2) The collection of self-administered data presents several possible biases such as social desirability bias, memory bias, methodological bias with filling fatigue among respondents, and cognitive bias due to ignorance of products consumed and altered memory capacities of many of the NPS.34 (3) The volatile nature of cocaine could partly explain the higher level of cocaine detected in hair versus questionnaires. Thus, a participant with a low level of cocaine detected once in the hair may not have necessarily used cocaine.

The overall prevalence of NPS use detected in hair in this MSM population (27/69, 39%) is higher than those reported in the literature using hair analysis in other types of population, such as nightclubs attendees, amphetamine or ketamine users, drivers under influence of drugs, acutely intoxicated patients in intensive care unit or drug-dependent patients, in which prevalence ranged between 7.8% and 32.5%,18,35 showing the high risk taken by this population. Our high prevalence may suggest a lack of knowledge among users of the drugs actually consumed, a misconception about the true composition of the products purchased most often on the net, and likely the use of NPS instead of ecstasy (MDMA) without user knowledge. This is of particular concern because of the possible consequences on somatic and neuropsychic health; these drugs potentially causing mental health problems, including risk for self-harm, anxiety, depression, and strong withdrawal effects as craving.6,9,36,37 However, the study population remains too small to demonstrate a significant difference in observed rates of psychiatric events between consumers and nonconsumers. In addition, psychiatric adverse events were not specifically requested at each visit and study physicians were not psychiatrists.

The overall prevalence of therapeutic drugs usage is high (50/69, 72%) with important rates of erectile drugs, ephedrine, and analgesic drugs (tramadol and codeine) consumption. Ephedrine is often found in methamphetamine or MDMA powders because it is a precursor of these drugs. Tramadol and codeine could be used as drug cutting agents. They pose many potential risks, especially because the correlation between consumer perceptions and test results is often low. The side effects of many drugs used as cutting agents are potentiated by alcohol, which may also increase and/or modify the effects of the drugs.

In our study, drug use is not associated with a specific sociodemographic profile in this high-risk, 30-year-old MSM population with a high school education and large majority of them working. Studies have shown that this use is perceived by most MSM as normal behavior, and this normalization is explained by the perceived pervasiveness of drug use.38

As already reported in the literature,14,17,32,39–41 drug use in our substudy is associated with risky sexual behaviors. Thus, drug users in this MSM population under PrEP have a higher number of partners in the past 2 months, more often casual partners, with more unprotected sex, more hard practices, and therefore have a higher risk of STIs confirmed in our study by an increased number of bacterial STIs in drug consumers during follow-up.

In our study, the comparison between drug use data from hair analysis and drug use data from questionnaire analysis showed that hair analysis detects more conventional drug and/or NPS use (P < 0.05). Hair analysis has an extended detection window of approximately 1 month per 1 cm of hair. Thus, a 1.5 cm length of hair captures a 90-day window of drug use. This detection window makes hair testing particularly attractive in case of (1) intermittent consumption, (2) absence of certainty as to the real composition of the consumed products, (3) absence of help of drug users by self-support group, and (4) in case of drug-induced memory impairment. The absence of removable hair can be circumvented by the use for testing of different matrices such as pubic or armpits hair or nails such as discussed in the consensus statement by the American Society of Addiction Medicine.42 Further studies must integrate this tool.

Although hair testing cannot be considered as a panacea for drug detection, it may also be used in combination with self-questionnaires to confirm reported abstinence.43 In the event that a patient's self-reported substance use differs from the results of a drug test, the physician could use this discrepancy as a springboard for therapeutic discussions.42

Limitations and Bias

This study had biases and limitations: (1) The study population represents only voluntary MSM with high-risk HIV behaviors taking PrEP and is not representative of the general MSM population. (2) The capillary matrix used for drug detection has limitations. Many participants had hair too short to be collected and analyzed. Pubic or axillary hair should have been sampled, but we did not include this possibility of a matrix in the consent signed by the participants. (3) The small size of our population limits the statistical power. (4) The late implementation of this substudy in the main trial resulted in some limitations. First, most hair analyses were performed on samples collected before the start of this substudy and did not allow sufficient hair to be collected to perform the GHB/GBL analysis in addition to other drugs. GHB/GBL remains a very frequently reported drug in this MSM population17,41 and in our study. Second, the list of drugs used by the self-administered questionnaires at each visit was extended from July 2015 with the addition of the NPS, crystal/methamphetamine, and speed/amphetamine classes, which did not allow a correlation to be established before that date between the drugs detected in the hair and those reported in the questionnaires at the same time, which would have made possible to analyze the possible amnesia effect of certain drugs and the quality control of the drugs consumed. Finally, the Committee for the Protection of Persons had requested that results not be communicated individually to the participants, removing any possibility of analyzing the impact in terms of risk reduction of the announcement of actual consumption over time associated with counseling according to each product. Hair testing is a promising and objective method of testing but not without limitations (eg, variable hair availability/length, availability of the technology, participant concerns about cosmetic visibility of sample collection, and higher relative cost). Hair testing also has several properties that make it potentially well suited for individuals with intermittent or lower frequency of use.


The high rate of drug use, particularly of NPS, raises questions about the somatic and psychological risks taken by MSM already at high risk of HIV infection and STIs. Self-reported drug use by questionnaires remains the reference tool for harm reduction at the individual level because of its feasibility and low cost. However, hair analysis is more sensitive with more molecules detected, especially molecules unknown to their users, gives an objective measurement of consumption, is reproducible, and is interesting to understand the uses and to be able to respond on a collective scale to risky practices with adapted messages, thus offering this group of patients a targeted multidisciplinary approach that includes psychologists, sexologists, addictionologists, and sexual health physicians.


The authors thank the participants of this study and the Fondation Pierre Bergé pour la prevention/Sidaction for a grant for the toxicological analysis.


1. Debruyne D, Le Boisselier R. Emerging drugs of abuse: current perspectives on synthetic cannabinoids. Subst Abuse Rehabil. 2015;6:113–129.
2. European Drug Report 2018: Trends and Developments EMCDDA. Lisbon, Portugal: EMCDDA; 2018.
3. Brunt TM, Poortman A, Niesink RJM, et al. Instability of the ecstasy market and a new kid on the block: mephedrone. J Psychopharmacol Oxf Engl. 2011;25:1543–1547.
4. Carhart-Harris RL, King LA, Nutt DJ. A web-based survey on mephedrone. Drug Alcohol Depend. 2011;118:19–22.
5. Zaami S, Giorgetti R, Pichini S, et al. Synthetic cathinones related fatalities: an update. Eur Rev Med Pharmacol Sci. 2018;22:268–274.
6. Weinstein AM, Rosca P, Fattore L, et al,. Synthetic cathinone and cannabinoid designer drugs pose a major risk for public health. Front Psychiatry. 2017;8:156.
7. Liakoni E, Yates C, Dines AM, et al. Acute recreational drug toxicity: comparison of self-reports and results of immunoassay and additional analytical methods in a multicenter European case series. Medicine (Baltimore). 2018;97:e9784.
8. Karila L, Megarbane B, Cottencin O, et al. Synthetic cathinones: a new public health problem. Curr Neuropharmacol. 2015;13:12–20.
9. Karila L, Billieux J, Benyamina A, et al. The effects and risks associated to mephedrone and methylone in humans: a review of the preliminary evidences. Brain Res Bull. 2016;126:61–67.
10. Batisse A, Grégoire M, Marillier M, et al. Cathinones use in Paris [in French]. L'Encephale. 2016;42:354–360.
11. Bourne A, Reid D, Hickson F, et al. “Chemsex” and harm reduction need among gay men in South London. Int J Drug Policy. 2015;26:1171–1176.
12. Melendez-Torres GJ, Bourne A. Illicit drug use and its association with sexual risk behaviour among MSM: more questions than answers? Curr Opin Infect Dis. 2016;29:58–63.
13. Bourne A, Reid D, Hickson F, et al. Illicit drug use in sexual settings (“chemsex”) and HIV/STI transmission risk behaviour among gay men in South London: findings from a qualitative study. Sex Transm Infect. 2015;91:564–568.
14. Sewell J, Miltz A, Lampe FC, et al. Poly drug use, chemsex drug use, and associations with sexual risk behaviour in HIV-negative men who have sex with men attending sexual health clinics. Int J Drug Policy. 2017;43:33–43.
15. Bracchi M, Stuart D, Castles R, et al. Increasing use of “party drugs” in people living with HIV on antiretrovirals: a concern for patient safety. AIDS Lond Engl. 2015;29:1585–1592.
16. McCall H, Adams N, Mason D, et al. What is chemsex and why does it matter? BMJ. 2015;351:h5790.
17. Roux P, Fressard L, Suzan-Monti M, et al. Is on-demand HIV pre-exposure prophylaxis a suitable tool for men who have sex with men who practice chemsex? Results from a substudy of the ANRS-IPERGAY trial. J Acquir Immune Defic Syndr. 2018;79:e69–e75.
18. Larabi IA, Fabresse N, Etting I, et al. Prevalence of new psychoactive substances in Paris ans its suburbs using hair analysis: a cross sectional study. Drug and Alcohol Dependence. 2019;204:107508.
19. Larabi A, Martin M, Fabresse N, et al. Hair testing for 3-fluorofentanyl, methoxyacetylfentanyl, carfentanil, acetylfentanyl and fentanyl by LC/MS/MS after unintentional overdose. Forensic Toxicol. 2020;38:277–286.
20. Tabb ZJ, Mmbaga BT, Gandhi M, et al. Antiretroviral drug concentrations in hair are associated with virologic outcomes among young people living with HIV in Tanzania. AIDS. 2018;32:1115–1123.
21. Molina JM, Capitant C, Spire B, et al. On-demand preexposure prophylaxis in men at high risk for HIV-1 infection. N Engl J Med. 2015;373:2237–2246.
22. Molina JM, Charreau I, Spire B, et al. Efficacy, safety, and effect on sexual behaviour of on-demand pre-exposure prophylaxis for HIV in men who have sex with men: an observational cohort study. Lancet HIV. 2017;4:e402–e410.
23. Pialoux G, Delaugerre C, Cotte L, et al. Pre-exposure prophylaxis: a useful tool to prevent human immunodeficiency virus infection? Clin Microbiol Infect. 2016;22:757–767. Review.
24. Fabresse N, Larabi IA, Stratton T, et al. Development of a sensitive untargeted LC-HRMS screening devoted to hair analysis through a shared MS2 spectra database: a step towards early detection of new psychoactive substances. Drug Test Anal. 2019;11:697–708.
25. Alvarez JC, Etting I, Abe E, et al. Identification and quantification of 4-methylethcathinone (4-MEC) and 3,4-methylenedioxypyrovalerone (MDPV) in hair using LC-MS/MS after chronic administration. Forensic Sci Int. 2017;270:39–45.
26. Alvarez JC, Lasne L, Etting I, et al. Hair analysis do not allow to discriminate between acute and chronic administration of a drug in young children. Int J Leg Med. 2018;132:165–172.
27. McCormack S, et al. Pre-exposure prophylaxis to prevent the acquisition of HIV-1 infection (PROUD): effectiveness results from the pilot phase of a pragmatic open-label randomised trial. Lancet Lond Engl. 2016;387:53–60.
28. Dolling DI, Desai M, McOwan A, et al. An analysis of baseline data from the PROUD study: an open-label randomised trial of pre-exposure prophylaxis. Trials. 2016;17:163.
29. Choopanya K, Martin M, Suntharasamai P, et al. Antiretroviral prophylaxis for HIV infection in injecting drug users in Bangkok, Thailand (the Bangkok Tenofovir Study): a randomised, double-blind, placebo-controlled phase 3 trial. Lancet. 2013;381:2083–2090.
30. Saxton P, Newcombe D, Ahmed A, et al. Illicit drug use among New Zealand gay and bisexual men: prevalence and association with sexual health behaviours. Drug Alcohol Rev. 2018;37:180–187.
31. HIV prevalence estimate among men who have sex with men attending gay venues in five French cities—PREVAGAY 2015. Available at: Accessed July 23, 2018.
32. Tomkins A, George R, Kliner M. Sexualised drug taking among men who have sex with men: a systematic review. Perspect Public Health. 2019;139:23–33.
33. Sewell J, Cambiano V, Miltz A, et al. Changes in recreational drug use, drug use associated with chemsex, and HIV-related behaviours, among HIV-negative men who have sex with men in London and Brighton, 2013–2016. Sex Transm Infect. 2018;94:494–501.
34. Palamar JJ, Salomone A, Vincenti M, et al. Detection of “bath salts” and other novel psychoactive substances in hair samples of ecstasy/MDMA/“Molly” users. Drug Alcohol Depend. 2016;161:200–205.
35. Alvarez JC, Fabresse N, Larabi IA. Prevalence and surveillance of synthetic cathinones use by hair analysis: an update review. Curr Pharm Des. 2017;23:1–9.
36. White CM. Mephedrone and 3,4-methylenedioxypyrovalerone (MDPV): synthetic cathinones with serious health implications. J Clin Pharmacol. 2016;56:1319–1325.
37. Santos GM, Coffin P, Santos D, et al. Feasibility, acceptability, and tolerability of targeted naltrexone for nondependent methamphetamine-using and binge-drinking men who have sex with men. J Acquir Immune Defic Syndr. 2016;72:21–30.
38. Ahmed AK, Weatherburn P, Reid D, et al. Social norms related to combining drugs and sex (“chemsex”) among gay men in South London. Int J Drug Policy. 2016;38:29–35.
39. Hirshfield S, Schrimshaw EW, Stall RD, et al. Drug use, sexual risk, and syndemic production among men who have sex with men who engage in group sexual encounters. Am J Public Health. 2015;105:1849–1858.
40. McCarty-Caplan D, Jantz I, Swartz J. MSM and drug use: a latent class analysis of drug use and related sexual risk behaviors. AIDS Behav. 2014;18:1339–1351.
41. Edmundson C, Heinsbroek E, Glass R, et al. Sexualised drug use in the United Kingdom (UK): a review of the literature. Int J Drug Policy. 2018;55:131–148.
42. Available at: Accessed November 4, 2020.
43. Gryczynski J, Schwartz RP, Mitchell SG, et al. Hair drug testing results and self-reported drug use among primary care patients with moderate-risk illicit drug use. Drug Alcohol Depend. 2014;141:44–50.

cathinones; new psychoactive substances; chemsex; hair; PrEP; HIV; MSM; harm reduction; addiction

Copyright © 2020 Wolters Kluwer Health, Inc. All rights reserved.