Characterization of People Living with HIV Who Inject Drugs in Mexico City: Importance for Transmission and Detection

Abstract

People who inject drugs (PWID) face a heightened risk of acquiring/transmitting HIV, enhanced by stigma and limited health care access. In Mexico, studies on PWID have focused on the north of the country. This study aimed to delineate characteristics of PWID living with HIV in Mexico City, identify profiles based on the substance injected, and evaluate variables associated with forming transmission clusters. A cross-sectional study was conducted with data from 2019 to 2023. Participants completed a questionnaire on sociodemographic, clinical, and behavioral variables. Bivariate and multi-variate logistic regression analyses were made. Among PWID, 96.3% were male (n = 437), of which 90.1% were men who have sex with men, 1.5% were cisgender females (n = 7), and 2.2% were transgender females (n = 10). PWID were more likely to use drugs during sex (adjusted odds ratio [aOR] = 3.3, 95% confidence interval [CI]: 1.7–6.4, p < 0.001), have more sexually transmitted diseases (aOR = 1.7, 95% CI: 1.1–2.9, p = 0.035), and have less condom use (aOR = 0.5, 95% CI: 0.3–0.8, p = 0.002). The most frequently injected substance was crystal meth, and those who injected it were more likely to have syphilis (aOR = 2.9, 95% CI: 1.2–7.1, p = 0.021), use Grindr (aOR = 3.6, 95% CI: 1.5–8.9, p < 0.001), and engage in high-risk practices (aOR = 6.9, 95% CI: 2.1–22.7, p < 0.001) in the last 3 months. Those under 25 years (p = 0.002), recently infected (p < 0.001), and who practiced insertive anal sex (p < 0.001) were more likely to be part of a cluster. These findings, and the increasing use of crystal meth, underscore the critical need to implement targeted risk-reduction strategies for PWID living with HIV and to design interventions responsive to specific profiles associated with different substances, taking into account not only their risk practices but also protective behaviors such as HIV testing.

Introduction

People who inject drugs (PWID) face both behavioral and environmental risks, making them a key population vulnerable to acquiring HIV. ^1,2 These risks are associated not only with injection practices but also with a higher propensity for engaging in risky sexual behaviors, such as having multiple sexual partners, unprotected sex, and sex work. ^{3 –5} In Mexico, the prevalence of HIV among PWID has been reported to range between 2.5% and 5%, ¹ with more than 80% contracting the hepatitis C virus (HCV). ^6,7 Since 2020, treatment for HCV is freely available through the National Hepatitis C Elimination Program and includes access to screening and diagnostic testing. About half of those who obtained reactive results had consumed substances via injection. ^{8 –10} However, studies of injecting substance use have focused on the north of the country and the border with the United States, with little information on the situation in Mexico City and other areas of the country. ^11,12

The World Health Organization (WHO) recognizes injecting drug use, particularly methamphetamines, as an escalating public health concern. ¹³ In Mexico, the use of amphetamines, methamphetamines, ecstasy, and other stimulants surged by 218% from 2017 to 2022, now representing the most significant demand for treatment, accounting for 46.2% of substance use treatment cases. ¹⁴ Methamphetamine use is especially prevalent among men who have sex with men (MSM), often linked to sexual contexts and intravenous administration. The widespread use of methamphetamines is largely attributed to their affordability and accessibility, with the drug even available through dating apps. In sexual contexts, methamphetamine use is associated with prolonged sexual activity and heightened sensory experiences. ^3,15,16

Health consequences for PWID are severe, with a heightened risk of acquiring other sexually transmitted infections (STIs) such as syphilis, alongside physical and mental health comorbidities, including anxiety and depression. ^5,17 PWID also experience significant stigma and discrimination, which contribute to lower utilization of diagnostic and health care services. ^18,19 This reduced engagement with health services is associated with lower levels of adherence to antiretroviral treatment (ART) and retention in care, increasing the likelihood of disease progression and HIV transmission to sexual partners. ^20,21

The study of HIV transmission networks has proven to be a valuable tool for targeting public health interventions. One way in which transmission networks can be inferred is by evaluating the genetic distances between all possible pairs of viral sequences from participants within a specific epidemiological context, enabling the inference of genetic transmission networks and the identification of “clusters” that group viruses with minimal genetic variation. ^{22 –25} This approach allows for the identification of priority clusters for intervention and the recognition of variables contributing to increased HIV transmission, which can then be targeted effectively. ^22,26

Leveraging the study of HIV transmission networks being carried out in Mexico City, ²⁷ this research aimed to identify the characteristics of people living with HIV (PLWH) who inject drugs in Mexico City. The objectives included assessing potential behavioral and sociodemographic differences between PWID and people who do not inject drugs (PWNID), identifying profile differences based on the drug injected, and determining the characteristics of PWID associated with a greater likelihood of belonging to a cluster.

Methods

Design and participants

This study is part of the HIV pre-treatment drug resistance and transmission network research in Mexico City ²⁷ carried out by the Center for Research in Infectious Diseases at the National Institute of Respiratory Diseases, a WHO reference laboratory for HIV genotyping. The research was conducted at the largest HIV clinics in Mexico City in its two branches: the Clínica Especializada Condesa Cuauhtémoc and the Clínica Especializada Condesa Iztapalapa. These clinics, operated by the Ministry of Health, play a pivotal role in HIV diagnosis in Mexico City, accounting for 70–80% of new HIV diagnoses in the metropolitan area, including both insured and uninsured individuals, mostly from the metropolitan area of Mexico City (including Mexico City and the surrounding municipalities of the State of Mexico). All persons with a positive HIV test result, naive to ART, and persons with prior exposure to antiretrovirals, who had been lost to follow-up and returned to clinical care to restart ART, were invited to participate. These criteria were followed according to the WHO recommendations to conduct pre-treatment drug resistance surveys. ²⁸

In this cross-sectional study, individuals who were recently diagnosed with HIV or who returned to clinical care after being lost to follow-up between July 2019 and December 2023 were invited to participate. Participants underwent viral genotyping and completed a self-administered, computer-based questionnaire covering sociodemographic, clinical, substance use, and sexual behavior variables.

Procedure and questionnaire

After providing written informed consent, participants were invited to undergo viral genotyping and complete a self-administered, computer-based questionnaire. The questionnaire covered a range of topics, including sociodemographic variables (e.g., age, gender, marital status, place of residence, educational level, socioeconomic status); clinical variables (e.g., time of diagnosis, CD4 T-cell count, viral load at diagnosis, history of sexually transmitted infections [STIs]); substance use, both injected and non-injected, within the last 3 months; and risky sexual practices (e.g., condom use, use of dating apps, participation in sex work, and frequented venues).

Recent HIV infection was identified using antibody avidity tests following a previously described algorithm based on two other previously validated multi-test algorithms, including HIV antibody avidity tests plus HIV viral load and CD4⁺ T-cell count (Sedia Biosciences, Beaverton, OR). ^{29 –31} Finally, the local HIV transmission network was inferred from the viral genetic sequences, spanning the protease and reverse transcriptase, obtained using standard next-generation sequencing techniques, as previously described, using genetic distance threshold methods. ³² Clusters were defined as sequences with Tamura Nei 93 genetic distance ≤1.5%. The network was inferred using a locally adapted version of the tool HIV-TRACE. ³³

Data analysis

Frequencies and percentages were calculated to describe the population of PWID in Mexico City. Fisher’s exact test was used to identify differences in sociodemographic, clinical, and behavioral variables between PWID and PWNID. In addition, bivariate and multi-variable logistic regression models were built with 95% confidence intervals [CIs] to assess the odds related to being a PWID. Statistically significant variables and potential confounders from the bivariate analysis were included. Different models were constructed using a backward elimination methodology to select a subset of explanatory variables for the model. The best model was selected by evaluating fit using the Akaike and Bayesian information criteria. All analyses were performed using RStudio v.2024.04.1 + 748. Afterward, only the sample of PWID was used to compare variations based on the type of drug injected and to determine factors in PWID associated with cluster membership.

Ethical considerations

The study was reviewed and approved by the Ethics, Research Ethics, and Biosafety Committees of the National Institute of Respiratory Diseases (Project Code: E02-20). All participants provided written informed consent to participate in the study.

Results

Characterization of PLWH who inject drugs in Mexico City

Of the 9805 respondents who answered the question on injecting substance use, 454 reported having used injectable drugs, accounting for 4.6% (almost 25% had not answered the question; total sample = 13,043). The vast majority of PWID were men (96.3%), with cisgender women comprising 1.5% and transgender women 2.2%. Over half (57.7%) were aged 30 or older, most resided in Mexico City (83.9%), and the majority were single (85.3%). Regarding education, 53.6% had completed high school, and 43.1% held a university or postgraduate degree. Among PWID, 89.1% reported injecting crystal meth, 5.6% cocaine, and 4.1% opiates, with 20.7% self-reporting a recent increase in crystal meth use. In addition to injection, drugs were consumed orally (34.1%), anally (17.9%), snorted (42.8%), and smoked (67.3%). Notably, there was an increase in PWID over time: 3.3% in 2019, 4.2% in 2020, 3.8% in 2021, 4.9% in 2022, and 5.7% in 2023 (p = 0.002).

Most participants who inject drugs (79.1%) were ART naive, and more than 20% were restarters; almost 2% had a history of using pre-exposure prophylaxis, and 30% had used post-exposure prophylaxis at some point. About half (48.9%) had CD4 T-cell counts between 200 and 499 cells/μL, and the majority (77.9%) lacked social insurance. Recent HIV infection (within 130 days) was found in 9.9% of participants who inject drugs, while 23.2% were resistant to at least one antiretroviral drug at the time of the study. Also, half of these participants (47.9%) were part of a transmission cluster. Risky sexual behaviors were prevalent (80.9%), including unprotected sex (58.5%), sex without lubricant (35.8%), drug use during sex (68.3%), participation in threesomes/orgies (44.7%), double penetration (12.8%), fisting (9.9%), and bugchasing (the practice of intentionally seeking an HIV infection; 3.5%). Additionally, 10.7% reported paying for sex, while 19.6% received payment for sex. Regarding STIs, 46.1% self-reported a recent infection, with syphilis (26.3%), hepatitis B or C (8.9%), and gonorrhea (6%) being the most common. A significant proportion (74.8%) used dating apps, primarily Grindr (68.6%) and Twitter (X; 18.1%). Further, 54% frequented sexual encounter venues, including steam rooms/saunas (7.4%), orgy sites (9.8%), cabins (7.9%), and, most notably, sex hotels (40.1%).

Differences between people who inject drugs and those who do not

Compared with individuals who do not inject drugs, PWID were significantly more likely to be male (96.3% vs. 91.3%, p < 0.001), single (85.3% vs. 79.4%, p = 0.024), older than 30 years (57.7% vs. 48.5%, p < 0.001), and reside in Mexico City (83.9% vs. 71.0%, p < 0.001). A higher proportion of PWID lacked social insurance (77.9% vs. 68.9%, p < 0.001), although they had better CD4 counts at diagnosis (>200 cells/mm³; 60.0% vs. 51.2%, p < 0.001) and were more likely to have been tested for HIV 1–2 times per year (42.0% vs. 31.2%, p < 0.001). They were also more likely to have returned to care after being lost to follow-up (26.7% vs. 6.4%, p < 0.001) and be restarting treatment (20.9% vs. 6.7%).

PWID were significantly more likely to have used any substance (through any route of administration) in the last 3 months (82.9% vs. 56.4%, p < 0.001), including tobacco (39.9% vs. 25.6%, p < 0.001), alcohol (44.7% vs. 38.5%, p = 0.012), marijuana (36.9% vs. 17.9%, p < 0.001), cocaine (20.7% vs. 7.7%, p < 0.001), crack/rock (8.8% vs. 1.9%, p < 0.001), poppers (49.5% vs. 18.6%, p < 0.001), Gamma-Hydroxybutyric acid (GHB) (19.4% vs. 1.1%, p < 0.001), and sildenafil (34.1% vs. 4.6%, p < 0.001), among others.

PWID reported a higher incidence of STIs (overall: 46.1% vs. 29.4%, p < 0.001), including syphilis (26.3% vs. 12.6%, p < 0.001), gonorrhea (6.0% vs. 2.7%, p < 0.001), hepatitis B or C (8.8% vs. 1.4%, p < 0.001), and urinary tract infections (11.7% vs. 7.8%, p = 0.006). Additionally, they engaged in risky sexual behaviors more frequently (overall: 80.9% vs. 62.4%, p < 0.001), such as fisting (9.9% vs. 2.9%, p < 0.001), double penetration (12.8% vs. 3.5%, p < 0.001), condomless sex (58.5% vs. 32.1%, p < 0.001), bugchasing (3.5% vs. 1.1%, p = 0.002), threesomes/orgies (44.7% vs. 16.7%, p < 0.001), drug use during sex (68.3% vs. 28.7%, p < 0.001), paying for sex (10.7% vs. 3.5%, p < 0.001), and exchanging sex for money (19.6% vs. 5.5%, p < 0.001).

Further, PWID were more likely to use dating apps (74.8% vs. 56.8%, p < 0.001), particularly Grindr (68.6% vs. 44.7%, p < 0.001) and Horner (7.1% vs. 4.0%, p = 0.003), though they used Tinder less frequently (5.1% vs. 8.5%, p = 0.010). They also visited sexual encounter venues more often (54.7% vs. 32.6%, p < 0.001), including steam rooms/saunas (7.4% vs. 5.0%, p = 0.039), orgy sites (9.8% vs. 1.9%, p < 0.001), and encounter hotels (40.1% vs. 18.6%, p < 0.001). All differences between PWID and those who do not inject drugs are detailed in Table 1 and Fig. 1.

Table 1.

Differences Between People Who Inject Drugs and People Who Do Not Inject Drugs

	PWN ID		PWID
	n	%	n	%	p
Sociodemographic characteristics
Gender
Cis man	8359	91.3	437	96.3	<0.001 **
Cis woman	606	6.5	7	1.5
Trans woman	201	2.1	10	2.2
Trans man	4	0.1	0	0.0
Age
<25 years	2271	24.3	69	15.2	<0.001 **
25–29 years	2545	27.2	123	27.1
30–35 years	2307	24.7	133	29.3
35 years	2220	23.8	129	28.4
Education
None/elementary	398	4.3	15	3.3	0.097
High school	5381	57.6	243	53.6
University/Postgraduate degree	3559	38.1	195	43.1
Place of residence
Mexico City	6634	71.0	381	83.9	<0.001 **
State of Mexico	2439	26.1	57	12.6
Other	273	2.9	16	6.5
Marital status
Unmarried	1418	15.3	45	10.0	0.024*
Married	369	4.0	16	3.6
Single	7343	79.4	382	85.3
Widowed	37	0.4	2	0.4
Other	85	0.9	3	0.7
Partner’s HIV status
Without HIV	1283	16.0	52	13.4	0.001 *
Living with HIV	1605	19.9	110	28.4
Unknown	3244	4.3	131	33.7
Single	1913	23.8	95	24.5
Traveling (last 6 months)
National	1565	37.2	79	42.0	0.307
International	239	5.7	12	6.4
No	2400	57.1	97	51.6
Self-perceived social class
Low	3821	41.1	174	38.4	0.382
Medium	5436	58.5	276	60.9
High	42	0.4	3	0.7
Year of participation
2019	89	96.7	3	3.3	0.002 *
2020	1267	95.8	56	4.2
2021	3108	96.2	124	3.8
2022	2811	95.1	146	4.9
2023	2076	94.3	125	5.7
Clinical characteristics
HIV test motivation
Routine checkup	1031	18.1	26	9.7	<0.001 **
Diagnosed partner	523	9.2	18	6.0
Recent risky practice	1215	21.4	69	23.0
Change in social security status	23	0.4	8	2.7
Treatment initiation	2106	37.3	91	30.3
Resumption of treatment after interruption and lost-to-follow-up	365	6.4	71	26.7
Other	425	7.5	17	5.6
Antiretroviral exposure
Naive	5208	93.2	234	79.1	<0.001 **
Restarter	368	6.7	62	20.9
Other	9	0.1	—	—
CD4+ T-cell count (cells/µL)
200	4560	48.8	181	40.0	<0.001 **
200–499	4151	44.5	221	48.9
≥500	626	6.7	50	11.1
Social insurance a
With social insurance	2465	26.5	69	15.2	<0.001 **
Without social insurance	6418	68.9	353	77.9
Unknown	427	4.6	31	6.8
ART resistance (any drug class) b
Yes	1927	21.2	103	23.2	0.313
Recent infection
Yes	773	8.3	45	9.9	0.223
Being in a cluster
Yes	4838	53.4	211	47.9	0.025 *
HIV testing frequency
First test	2069	36.6	94	21.3	<0.001 **
1–2 times a year	1767	31.2	126	42.1
1 time every 2 years or less	1466	25.9	79	26.3
If a risky practice occurs	359	6.3	31	10.3
Any STI c
No	5710	65.2	200	46.1	<0.001 **
Yes	2578	29.4	200	46.1
Unknown	472	5.4	34	7.8
Behavioral characteristics (last 3 months)
Substance use d
Yes	4910	56.4	360	82.9	<0.001 **
High risk behavior e
Yes	3272	62.4	228	80.9	<0.001 **
Paid for sex
Yes	267	3.5	41	10.7	<0.001 **
Received payment for sex
Yes	311	5.5	59	19.6	<0.001 **
Role in anal sex
Receptive	2728	32.3	120	28.6	0.014 *
Insertive	1759	20.8	71	17.0
Versatile	3536	41.9	209	49.9
Mostly receptive	425	5.0	19	4.5
Dating apps use
Yes	4885	56.8	326	74.8	<0.001 **
Attending sexual encounter venues
Yes	2710	32.6	228	54.7	<0.001 **

^aThe Mexican health system is administratively fragmented. Changes in social security status (tightly linked to employment status) imply a change of HIV clinic, as persons without social security are cared for at Ministry of Health clinics and persons with social security are cared for at clinics run by social security institutions. This change of systems often implies repeating the HIV diagnostic test.

^bAntiretroviral drug resistance was assessed from pol sequences, using the Stanford HIVdb tool. If the Stanford score was >15 for any drug, the virus was considered resistant.

^cSTI: sexually transmitted infections during the past 3 months; queried as syphilis, gonorrhea, human papillomavirus, urinary tract infection, hepatitis, and others. Self-reported. Individuals could select one or more of the options.

^dSubstance use: asked as tobacco, alcohol, marijuana, cocaine, poppers, amphetamines, hallucinogens, crystal meth, GHB, or others. Self-reported. Individuals could select one or more of the options.

^eHigh-risk behavior including fisting, double penetration, lubricant-free sex, condomless sex, bug chasing, threesomes/orgies, chemsex, and use of drugs during sex. Self-reported. Individuals could select one or more of the options.

^* p < 0.05.

^** p < 0.001.

PWID, people who inject drugs; PWNID, people who do not inject drugs; STI, sexually transmitted infection.

FIG. 1.

Comparisons between people who inject and people who do not inject drugs and depending on the drug injected. HBV or HCV, hepatitis B or C virus; HPV, human papillomavirus; PWID, people who inject drugs; PWNID, people who do not inject drugs; STI, sexually transmitted infections.

Correcting for confounding variables, those with an STI diagnosis in the last 3 months (adjusted odds ratio [aOR] = 1.7, 95% CI: 1.1–2.9, p = 0.035), who used drugs during sex in the last 3 months (aOR = 3.3, 95% CI: 1.7–6.4, p < 0.001), and who have increased their use of crystal meth in the last 3 months (aOR = 5.4, 95% CI: 2.3–12.8, p < 0.001) were more likely to inject drugs. On the contrary, those who used condoms more than 50% of the time (aOR = 0.5, 95% CI: 0.3–0.8, p = 0.002), and those who were part of a cluster (aOR = 0.6, 95% CI: 0.4–0.9, p = 0.037), were less likely to be PWID (Table 2).

Table 2.

Logistic Regression Model on Injecting Substance Use

Variable	Category	OR	95% CI	p-Value	aOR a	95% CI	p-Value
Gender	Men	Reference			Reference
	Women	0.23	0.11–0.48	<0.001 *	1.38	0.21–8.99	0.734
	Trans women	0.98	0.52–1.87	0.960	0.86	0.17–4.50	0.861
Age	<25	Reference			Reference
	25–29	1.62	1.19–2.18	<0.001 *	0.77	0.40–1.48	0.433
	30–35	1.91	1.42–2.58	<0.001 *	0.85	0.43–1.67	0.643
	35 or more	1.91	1.41–2.58	<0.001 *	1.29	0.63–2.64	0.491
Being part of a cluster		0.79	0.65–0.96	0.018 *	0.60	0.37–0.97	0.037 *
CD4+ cell count (cells/mm3)	<200	Reference			Reference
	200–500	1.34	1.09–1.64	0.005 *	1.01	0.60–1.68	0.983
	>500	2.07	1.50–2.86	<0.001 *	1.07	0.43–2.62	0.891
Sex partners last 3 months	1	Reference			Reference
	2–5	2.01	1.51–2.69	<0.001 *	0.80	0.38–1.69	0.565
	6–20	3.05	2.25–4.15	<0.001 *	1.20	0.37–3.87	0.762
	21–50	3.33	2.13–5.20	<0.001 *	0.57	0.09–3.80	0.563
	51–100	2.98	1.77–5.00	<0.001 *	0.66	0.27–1.61	0.360
	Over 100	5.75	3.14–10.53	<0.001 *	1.88	0.35–10.12	0.461
Sexual partners via dating app	<4	Reference			Reference
	4 or more	2.21	1.73–2.83	<0.001 *	1.66	0.96–2.84	0.068
Partner living with HIV	Yes	Reference			Reference
	No	0.62	0.44–0.86	0.010 *	1.41	0.64–3.08	0.392
	Unknown	0.61	0.47–0.80	<0.001 *	0.82	0.45–1.53	0.539
	No partner	0.75	0.57–1.00	0.050	–	–	–
Previous ART use		2.23	1.24–4.01	0.010 *	0.36	0.04–3.15	0.359
Condon use >50%		0.50	0.38–0.65	<0.001 *	0.45	0.27–0.74	0.002 *
STI last 3 months b	No	Reference			Reference
	Yes	2.23	1.82–2.73	<0.001 *	1.73	1.04–2.89	0.035 *
	Unknown	2.01	1.38–2.94	<0.001 *	2.49	1.18–5.28	0.017 *
Drug use during sex last 3 months		6.83	5.38–8.65	<0.001 *	3.33	1.73–6.39	<0.001 *
GHB use last 3 months		20.92	15.35–28.49	<0.001 *	1.76	0.70–4.42	0.233
Sildenafil use last 3 months		10.94	8.75–13.67	<0.001 *	1.72	0.93–3.19	0.083
Crystal use increase last 3 months	No	Reference			Reference
	Yes	17.93	12.10–26.57	<0.001 *	5.44	2.31–12.78	<0.001 *
	Not applicable	0.23	0.14–0.37	<0.001 *	0.70	0.28–1.74	0.442
Pay for sex in the las 3 months		3.36	2.38–4.76	<0.001 *	1.01	0.36–2.79	0.991
Received pay for sex		4.26	3.14–5.79	<0.001 *	1.22	0.60–2.47	0.578
Last risk sexual practice	Never	Reference			Reference
	<2 weeks	2.26	1.32–3.87	<0.001 *	1.11	0.12–10.60	0.928
	2–4 weeks	2.39	1.41–4.06	<0.001 *	3.85	0.47–31.48	0.209
	1–3 months	2.23	1.38–3.60	<0.001 *	4.19	0.53–32.90	0.173
	>3 months	1.81	1.14–2.87	<0.001 *	5.54	0.71–43.02	0.101

^aAdjusted odds ratio (aOR). All the variables shown in the table were considered as potential confounders.

^bSTI: sexually transmitted infections during the past 3 months; queried as syphilis, gonorrhea, human papillomavirus, urinary tract infection, hepatitis, and others. Self-reported individuals could select one or more of the options.

^* p < 0.05.

ART, antiretroviral treatment; CI, confidence interval; OR, odds ratio.

Differences based on the drug injected

In the sample of PWID (n = 454), compared with individuals who injected other substances, those who injected crystal meth were significantly more likely to be male (98.9% vs. 88.3%, p = 0.002), with very few cisgender or transgender women reporting crystal meth use (n = 4, 1.1%). People who injected crystal meth were also more likely to have a university degree compared to those who injected cocaine or opiates (48.7% vs. 23.8%, p < 0.001) and less likely to report being from a lower social class (34.8% vs. 57.1%, p = 0.024). Further, they were more likely to undergo periodic HIV testing (1 or 2 times per year; 44.3% vs. 28.6%, p = 0.022). In terms of substance use, people who injected crystal meth were more likely to have used other substances in the past 3 months (85.5% vs. 64.3%, p = 0.002). They reported higher use of poppers (54.8% vs. 11.9%, p < 0.001), GHB (22.9% vs. 0.0%, p < 0.011), and sildenafil (38.5% vs. 4.8%, p < 0.001). They also reported an increase in alcohol consumption in recent months (89.7% vs. 70.8%, p = 0.016).

People who injected crystal meth engaged in risky behaviors more frequently overall (88.0% vs. 42.3%, p < 0.001). This included more instances of sex without lubricant (38.3% vs. 15.4%, p = 0.029), sex without a condom (65.4% vs. 15.4%, p < 0.001), participation in group sex (50.7% vs. 11.5%, p < 0.001), and drug use during sex (72.9% vs. 36.7%, p < 0.001). They also reported higher use of dating apps (79.8% vs. 47.5%, p < 0.001), particularly Grindr (74.6% vs. 32.5%, p < 0.001) and Twitter (X; 19.9% vs. 2.5%, p = 0.004), and more frequent visits to venues, especially hotels (42.3% vs. 23.7%, p = 0.035). Differences based on the injected drug are detailed in Table 3 and Fig. 1.

Table 3.

Differences Depending on the Injected Substance

	Inject other		Inject crystal
	n	%	n	%	p
Sociodemographic characteristics
Gender
Cis man	38	88.3	349	98.9	0.002 *
Cis woman	2	4.7	1	0.3
Trans woman	3	7.0	3	0.8
Trans man	—	—	—	—
Age
<25 years	9	20.9	55	15.6	0.068
25–29 years	8	18.6	100	28.3
30–35 years	8	18.6	106	30.0
>35 years	18	41.9	92	26.1
Education
None/elementary	5	11.9	7	2.0	<0.001 **
High school	27	64.3	174	49.3
University/Postgraduate degree	10	23.8	172	48.7
Place of residence
Mexico City	36	83.7	298	84.4	0.933
State of Mexico	6	14.0	42	11.9
Other	1	2.3	13	3.7
Marital status
Unmarried	5	11.7	29	8.3	0.520
Married	3	7.0	11	3.2
Single	35	81.3	303	8 7.3
Widowed	—	—	2	0.6
Other	—	—	2	0.6
Partner’s HIV status
Without HIV	9	28.1	36	11.7	0.110
Living with HIV	8	25.1	87	28.2
Unknown	9	29.1	104	33.8
Single	6	18.7	81	26.3
Traveling (last 6 months)
National	5	31.2	66	42.9	0.458
International	—	—	10	6.5
No	11	68.8	78	50.6
Self-perceived social class
Low	24	57.1	123	34.8	0.024 *
Medium	18	42.9	228	64.6
High	—	—	2	0.6
Year of participation
2019	—	—	3	100.0	0.430
2020	9	17.0	44	83.0
2021	12	11.0	97	89.0
2022	15	11.4	117	88.6
2023	7	7.1	92	92.9
Clinical characteristics
HIV test motivation
Routine checkup	3	10.0	19	8.4	0.865
Diagnosed partner	2	6.7	12	5.3
Recent risky practice	6	20.0	56	24.7
Change in social security status	—	—	5	2.2
Treatment initiation	10	33.3	66	29.1
Resumption of treatment after interruption and lost-to-follow-up	6	20.0	57	25.1
Other	3	10.0	12	5.3
Antiretroviral exposure
Naive	26	89.7	173	77.2	0.152
Restarter	3	10.3	51	22.8
Other	—	—	—	—
CD4+ T-cell count (cells/µL)
<200	22	51.2	135	38.3	0.146
200–499	15	34.9	177	50.3
≥500	6	13.9	40	11.4
Social insurance a
With social insurance	11	25.6	53	15.1	0.244
Without social insurance	30	69.8	278	79.0
Unknown	2	4.6	21	5.9
ART resistance (any drug class) b
Yes	10	23.8	87	25.1	1.000
Recent infection
Yes	6	14.0	36	10.2	0.434
Being in a cluster
Yes	18	42.9	164	47.7	0.625
HIV test frequency
First test	12	42.9	40	17.5	0.022 *
1–2 times a year	8	28.6	101	44.3
1 time in 2 years or less	7	25.0	61	26.8
If a risky practice occurs	1	3.5	26	11.4
Any STI c
No	19	50.0	150	43.5	0.534
Yes	15	39.5	167	48.4
Unknown	4	10.5	28	8.1
Behavioral characteristics (last 3 months)
Substance use d
Yes	27	64.3	295	85.5	0.002 *
High risk behavior e
Yes	11	42.3	191	88.0	<0.001 **
Paid for sex
Yes	2	6.1	35	11.5	0.556
Received payment for sex
Yes	2	6.7	44	19.2	0.126
Role in anal sex
Receptive	5	13.1	101	30.2	0.003 *
Insertive	9	23.7	54	16.2
Versatile	18	47.4	168	50.3
Mostly receptive	6	15.8	11	3.3
Dating apps use
Yes	19	47.5	277	79.8	<0.001 **
Attending sexual encounter venues
Yes	15	39.5	187	56.2	0.059

^aThe Mexican health system is administratively fragmented. Changes in social security status (tightly linked to employment status) imply a change of HIV clinic, as persons without social security are cared for at Ministry of Health clinics and persons with social security are cared for at clinics run by social security institutions. This change of systems often implies repeating the HIV diagnostic test.

^bAntiretroviral drug resistance was assessed from pol sequences, using the Stanford HIVdb tool. If the Stanford score was >15 f or any drug, the virus was considered resistant.

^cSTI: sexually transmitted infections during the past 3 months; queried as syphilis, gonorrhea, human papillomavirus, urinary tract infection, hepatitis, and others. Self-reported. Individuals could select one or more of the options.

^dSubstance use: asked as tobacco, alcohol, marijuana, cocaine, poppers, amphetamines, hallucinogens, crystal meth, GHB, or others. Self-reported. Individuals could select one or more of the options.

^eHigh-risk behavior: including fisting, double penetration, lubricant-free sex, condomless sex, bug chasing, threesomes/orgies, chemsex, and use of drugs during sex. Self-reported. Individuals could select one or more of the options.

^* p < 0.05.

^** p < 0.001.

ART, antiretroviral treatment.

Correcting for confounding variables, those who inject crystal meth were more likely to have had syphilis in the last 3 months (aOR = 2.9, 95% CI: 1.2–7.1, p = 0.021), to have used Grindr in the past 3 months (aOR = 3.6, 95% CI: 1.5–8.9, p < 0.001), to have engaged in high-risk practices in the last 3 months (aOR = 6.9, 95% CI: 2.1–22.7, p < 0.001), and to have a negative HIV test at least 1 year prior (aOR = 3.3, 95% CI: 1.1–9.8, p < 0.001) (Table 4).

Table 4.

Logistic Regression Model on Injecting Crystal Meth Versus Injecting Other Substances

Variable	Category	OR	95% CI	p-Value	aOR a	95% CI	p-Value
Use of antiretroviral drugs	Never	Reference			Re
	Post-exposure prophylaxis (PEP)	0.54	0.18–1.62	0.272	1.77	0.17–18.35	0.632
	Pre-exposure prophylaxis (PrEP)	8.09	1.06–61.66	0.044 *	9.65	1.05–88.65	0.045*
	ART	1.67	0.87–3.24	0.126	0.96	0.14–6.72	0.966
Syphilis last 3 months		2.20	1.16–4.17	0.016 *	2.88	1.17–7.05	0.021 *
Poppers use last 3 months		2.94	1.77–4.86	<0.001 *	1.06	0.43–2.56	0.906
GHB use last 3 months		4.99	1.96–12.73	0.001 *	2.03	0.60–6.90	0.255
Sildenafil use last 3 months		3.10	1.71–5.61	<0.001 *	0.58	0.22–1.53	0.273
Grindr use last 3 months		3.61	2.23–5.84	<0.001 *	3.62	1.49–8.76	0.004 *
High risk behavior last 3 months b		5.56	2.93–10.54	<0.001 *	6.93	2.12–22.69	0.001 *
Drug use during sex last 3 months		2.37	1.41–3.98	0.001 *	0.39	0.13–1.19	0.098
Last risk sexual practice	Never	Reference			Reference
	<2 weeks	2.17	0.74–6.38	0.161	0.55	0.11–2.85	0.479
	2–4 weeks	2.58	0.89–7.52	0.082	0.74	0.14–3.84	0.721
	1–3 months	5.19	1.87–14.44	0.002 *	1.09	0.23–5.25	0.916
	>3 months	3.21	1.27–8.11	0.014 *	1.93	0.52–7.12	0.327
Last negative HIV self-reported test	<6 months	Reference			Reference
	6 months to 1 year	1.64	0.63–4.22	0.308	3.51	0.97–12.72	0.055
	Over a year	1.78	0.82–3.90	0.147	3.30	1.11–9.82	0.032 *
	First time	0.62	0.26–1.51	0.292	2.31	0.65–8.22	0.198
HIV test motivation	Routine	Reference			Reference
	Diagnosed partner	0.74	0.20–2.73	0.647	1.00	0.15–6.71	0.999
	Recent risk event	1.59	0.55–4.56	0.391	1.36	0.33–5.67	0.671
	Change in social security	0.61	0.12–3.27	0.568	0.31	0.02–4.19	0.380
	Treatment initiation	0.97	0.37–2.60	0.956	1.34	0.35–5.18	0.673
	Returning to care	1.50	0.53–4.27	0.447	2.28	0.24–21.71	0.473
	Other	0.88	0.23–3.43	0.859	2.07	0.28–15.43	0.477

^aAdjusted odds ratio (aOR). All the variables shown in the table were considered as potential confounders.

^bHigh-risk behavior: including fisting, double penetration, lubricant-free sex, condomless sex, bugchasing, threesomes/orgies, chemsex, and use of drugs during sex. Self-reported. Individuals could select one or more of the options.

^* p < 0.05

ART, antiretroviral treatment; CI, confidence interval; OR, odds ratio.

Characteristics of PWID associated with belonging to a cluster

Within the group of PWID, our analysis identified several characteristics associated with a higher likelihood of belonging to a transmission cluster. Those PWID under 25 years of age were significantly more likely to be in a cluster (63.7% vs. 39.8%, p = 0.002) than those who were older. Additionally, recent infection (73.3% vs. 45.1%, p < 0.001) and CD4 counts greater than 500 cells/mL (62.2% vs. 45.9%, p = 0.030) were associated with an increased probability of being in a cluster. PWID who used cocaine through any route of administration (58.8% vs. 44.0%, p = 0.021) or ketamine (70.0% vs. 45.9%, p = 0.040) in the past 3 months were also more likely to cluster. Further, involvement in chemsex (54.7% vs. 39.7%, p = 0.016) and having an insertive role in anal sex, as opposed to a receptive or versatile role (61.4% vs. 45.7%, p < 0.001), were linked to cluster membership (Table 5). Multi-variate analyses were performed, although no significant model was found.

Table 5.

Characteristics of PWID Associated with Belonging to a Cluster

	No cluster		Cluster
	n	%	n	%	p
Sociodemographic characteristics
Gender
Cis man	222	52.4	202	47.6	0.929
Cis woman	3	42.9	4	57.2
Trans woman	5	50.0	5	50.0
Age
<25 years	25	36.2	44	63.8	0.002 *
25–29 years	55	45.8	65	54.2
30–35 years	76	58.5	54	41.5
>35 years	74	60.7	48	39.3
Education
None/elementary	7	44.7	8	53.3	0.150
High school	114	48.1	123	51.9
University/Postgraduate degree	108	57.5	80	42.5
Place of residence
Mexico City	193	52.4	175	47.5	0.853
State of Mexico	28	49.1	29	50.9
Other	9	56.2	7	43.8
Marital status
Unmarried	21	47.7	23	52.3	0.413
Married	5	31.3	11	68.7
Single	197	53.2	173	46.8
Widowed	1	50.0	1	50.0
Other	1	33.3	2	66.7
Partner’s HIV status
Without HIV	27	51.9	25	48.1	0.211
Living with HIV	55	52.4	50	47.6
Unknown	58	45.3	70	54.7
Single	55	59.8	37	40.2
Traveling (last 6 months)
National	40	52.6	36	47.4	0.382
International	9	75.0	3	25.0
No	49	52.7	44	47.3
Self-perceived social class
Low	86	50.9	83	49.1	0.666
Medium	143	53.4	125	46.6
High	1	33.3	2	66.7
Year of participation
2019	2	66.7	1	33.3	0.716
2020	27	48.2	29	51.8
2021	65	54.6	54	45.4
2022	69	48.6	73	51.4
2023	67	55.4	54	44.6
Clinical characteristics
HIV test motivation
Routine checkup	9	36.0	16	64.0	0.067
Diagnosed partner	10	55.6	8	44.4
Recent risky practice	28	40.6	41	59.4
Change in social security status	3	60.0	2	40.0
Treatment initiation	43	48.3	46	51.7
Resumption of treatment after interruption and lost-to-follow-up	45	65.2	24	34.8
Other	10	58.8	7	41.2
Antiretroviral exposure
Naive	110	47.8	120	52.2	0.019 *
Restarter	38	65.5	20	34.5
CD4− T cells count (cells/µL)
<200	105	58.7	74	41.3	0.030 *
200–499	108	50.2	107	49.8
≥500	17	37.8	28	62.2
Social insurance a
With social insurance	28	41.8	39	58.2	0.168
Without social insurance	186	54.4	156	45.6
Unknown	16	51.6	15	48.4
ART resistance (any drug class) b
Yes	56	54.4	47	45.6	0.653
Recent infection
Yes	12	26.7	33	73.3	<0.001 **
HIV test frequency
First test	26	41.9	36	58.1	0.389
1–2 times a year	61	50.0	61	50.0
1 time in 2 years or less	44	56.4	34	43.6
If a risky practice occurs	16	53.3	14	46.7
Any STI c
No	102	51.8	95	48.2	0.582
Yes	103	53.9	88	46.1
Unknown	15	44.1	19	55.9
Behavioral characteristics (last 3 months)
Substance use d
Yes	178	51.0	171	49.0	0.096
High risk behavior e
Yes	111	49.8	112	50.2	0.126
Paid for sex
Yes	24	58.5	17	41.5	0.412
Received payment for sex
Yes	36	47.3	29	52.7	0.654
Role in anal sex
Receptive	57	49.6	58	50.4	<0.001 *
Insertive	27	38.6	43	61.4
Versatile	109	53.7	94	46.3
Mostly insertive	17	89.5	2	10.5
Dating apps use
Yes	163	51.9	151	48.1	0.825
Attending sexual encounter venues
Yes	109	49.5	111	50.5	0.371

^aThe Mexican health system is administratively fragmented. Changes in social security status (tightly linked to employment status) imply a change of HIV clinic, as persons without social security are cared for at Ministry of Health clinics and persons with social security are cared for at clinics run by social security institutions. This change of systems often implies repeating the HIV diagnostic test.

^bAntiretroviral drug resistance was assessed from pol sequences, using the Stanford HIVdb tool. If the Stanford score was >15 for any drug, the virus was considered resistant.

^cSTI: sexually transmitted infections during the past 3 months; queried as syphilis, gonorrhea, human papillomavirus, urinary tract infection, hepatitis, and others. Self-reported individuals could select one or more of the options.

^dSubstance use: asked as tobacco, alcohol, marijuana, cocaine, poppers, amphetamines, hallucinogens, crystal meth, GHB, or others. Self-reported. Individuals could select one or more of the options.

^eHigh-risk behavior: including fisting, double penetration, lubricant-free sex, condomless sex, bugchasing, threesomes/orgies, chemsex, and use of drugs during sex. Self-reported individuals could select one or more of the options.

^* p < 0.05.

^** p < 0.001.

ART, antiretroviral treatment; PWID, people who inject drugs.

Discussion

In the study population, over the past 5 years, more than 4% of individuals newly diagnosed with HIV in the Mexico City network reported injecting drugs. This percentage has increased annually, reaching nearly 6% in 2023, which is somewhat higher compared with similar populations in the United States. ³⁴ The majority of these individuals were male, and the predominant substance injected was crystal meth, consistent with findings from other studies on similar populations. ^{16,34 –37}

Our results are consistent with previous research that found a significant correlation between substance use and risky sexual practices. ³⁶ PWID were more likely to engage in high-risk behaviors, such as condomless sex, group sex, and chemsex, compared with those who did not use injectable drugs. They were also more likely to have a recent STI diagnosis and to participate in practices like fisting and sex work, which pose greater physical and sexual health risks and increase vulnerability to violence. ^{38 –42} This heightened risk is further underscored by the fact that nearly half of these individuals had recently been diagnosed with an STI. ^43,44 Understanding the interplay between substance use and sexual behavior is crucial for developing effective harm reduction and substance use interventions. Strategies should address the link between substances and sexual activity, helping individuals reconnect with their sexual desires while reducing the association between drug use and risky sexual behaviors. ⁴⁵ In addition, PWID were found to be more likely to be restarters, suggesting a relationship between substance use and loss to medical follow-up, which should be addressed in future studies and interventions. ⁴⁶

Among participants who inject drugs, it was observed that many used dating apps, especially Grindr, which can be used not only to meet people but specifically to find sexual partners who also inject drugs and maybe find substance providers. This behavior, prevalent across various populations, could be linked to negative self-concept, limited social skills, or a restricted social context, ^47,48 although these hypotheses need to be studied further to be confirmed in the MSM population. Additionally, there may be a connection between relieving anxiety or discomfort and engaging in sexual activity, which might explain the immediate use of apps to find sexual partners. ⁴⁹ Many also frequented specific venues for sexual encounters, despite the associated health and safety risks. ^{50 –52} This suggests that individuals who inject drugs and live with HIV in Mexico City display distinct patterns in their choice of meeting places and apps, which are often different from those used by individuals who do not inject drugs. These patterns could be targeted in risk reduction strategies or combined with mobile health interventions. ^45,53 Notably, certain hotels emerge as common locations for finding both sexual partners and drug dealers, reinforcing the link between drug use and sex. Interventions in these settings could be beneficial, as PWID are often stigmatized and face barriers to accessing health care services. ^{53 –55}

Interestingly, those who used injectable drugs had higher CD4⁺ T-cell counts at the time of the study and underwent diagnostic tests with certain frequency. This may indicate a higher perception of risk among PWID, which could be leveraged to implement preventive measures. These might include regular testing for HIV and other STIs, including the availability of tests in meeting venues, as well as enhancing retention and adherence to ART through mobile interventions. ⁵⁶ Notably, in the context of Mexico City, health, counseling, and diagnostic services focused on the LGBTI+ population provided by the Condesa clinic and its location in one of the gay-friendly neighborhoods of Mexico City may be important for access to regular HIV testing. Additionally, providing risk reduction workshops and awareness talks at the same locations where these screening tests are performed could equip individuals with the necessary tools to safeguard their health and well-being. ^45,57

Another approach that can be implemented in this regard is the use of post-exposure doxycycline to prevent sexually transmitted diseases (STDs), which has been shown to be effective for use among MSM and transgender women and can be well accepted and valued, ⁵⁸ although it is not yet available and has not yet been approved in Mexico. On the other hand, needle exchange programs or paraphernalia for consumption can be an effective strategy that should be implemented nationally and not only by civil associations, since so far such programs have been limited, especially to the northern border of the country, as is the case of Tijuana. ^59,60

Examining the differences in substance use among PWID, it was found that crystal meth was the most commonly used substance, probably related to its low cost and addictive potential, with a significant majority of PWID reporting its use, followed by cocaine and opiates in much lower proportions. People who injected crystal meth were predominantly male, with very few women or trans women reporting its use; in contrast, women and trans women represented 5% and 7% of those using other substances. This pattern supports the association of crystal meth use with sexual practices within the MSM group. Additionally, those using crystal meth tended to have higher educational attainment and potentially better economic status compared with people who use other substances. These characteristics, such as higher educational levels and greater use of apps, may bring benefits when implementing interventions remotely. ⁶¹ People who inject crystal meth also exhibited more STI diagnoses, engaged in risky sexual practices more often, and exhibited higher usage of dating apps and other substances, including GHB, sildenafil, and poppers. This reinforces the notion that crystal meth use in this population is often linked to sexual contexts. ^62,63

Further, these individuals had a higher perception of risk and underwent more regular HIV testing compared to those using other substances, who were more likely to have their first test. It is important to remember that these findings apply to PWID living with HIV who receive care in a clinic attended mostly by MSM and provide targeted services to this population. The situation in other types of populations, such as heterosexual men and women, must be studied in their own context. Even so, these observations underscore the need for tailored risk reduction and intervention strategies based on the specific substance used, given the distinct profiles and potentially differential access to health promotion services among people who use different drugs. ^45,64 It is also important to mention that the study contains data from the period of the COVID-19 pandemic, which may have influenced the number of people diagnosed with both HIV and STDs, since it has been widely reported, both in Mexico and in the world, that the number of tests and detections had a significant decrease in this period. In this sense, the results should be understood taking into consideration the context in which they were performed and taking into account the influence that the pandemic may have had. ^65,66

When analyzing these results, it is crucial to acknowledge the limitations and potential biases of these findings. First, although the sample was very large, data on substance use questions were not available for all participants, so perhaps the injection substance use may be underestimated. Also, a limited sample size or underrepresentation of certain groups could skew the findings, potentially underestimating or overestimating the actual connections within these populations. Specifically, people who receive the diagnosis in a poor state of health may be underrepresented in this study, since, globally, in Mexico they may account for up to 50% of diagnoses. ⁶⁷ Also, the study focuses only on one center. Even when the Condesa clinics centralize HIV diagnosis in Mexico City’s metropolitan zone, the inclusion of other clinics, mainly those providing services for the insured population, may be informative. Future research with more representative samples could be useful to validate these findings and elucidate the underlying mechanisms driving possible differences between key populations.

Additionally, the sociodemographic, behavioral, and clinical data were collected through a self-administered questionnaire, which, given the sensitivity of the topics, may lead to unreliable responses and introduce biases. The correlational nature of the study limits the ability to establish causal relationships, and these should be further explored in subsequent research. It is also important to remember that molecular networks are context-specific, and caution should be exercised when generalizing these results to other populations.

Regarding the implications of the study, it highlights that crystal meth use represents a significant health issue within the HIV-positive community in Mexico City, particularly among MSM and those who inject substances. This drug is associated with a higher frequency of risky sexual practices and increased health problems, and its prevalence is rising annually. Addressing this issue is urgent, with a focus on its relationship with access to health services, HIV diagnosis, prevention efforts (including specifically tailored pre-exposure prophylaxis programs, which have been available throughout Mexico since 2021 to attend, among others, to PWID), ⁶⁸ and adherence to ART. Notably, the relationship between crystal meth use and sex, especially in the MSM population, should not be ignored when designing interventions and harm reduction strategies. Interventions should be tailored to specific profiles and needs based on the substances used and the underlying motivations for use. Further, ongoing monitoring of transmission networks is essential to identify priority groups and implement targeted risk reduction interventions for this population, which faces multiple health challenges, including HIV and other STIs.

Authors’ Contributions

E.G.-V.: Conceptualization, data curation, formal analysis, research, project management, methodology, and writing—original draft. S.E.S.-M.: Data curation, formal analysis, research, and methodology. V.M.D.-C.: Conceptualization, project management, supervision, and writing—review and editing. N.P.C.-S.: Research, supervision, and writing—review and editing. S.Á.-R.: Conceptualization, acquisition of funds, research, project management, supervision, and writing—review and editing.

Data Availability

Due to the confidentiality and sensitivity of the information, the database is not shared publicly; for any queries about the data, the authors can be consulted.

Author Disclosure Statement

No potential conflict of interest was reported by the authors.

Funding Information

This work was supported by grants from Consejo Nacional de Humanidades Ciencia y Tecnología (CONAHCyT—PRONAII Virología 303079; and No. scholarship holder: 698506). The work was supported in part by funds from the San Diego Center for AIDS Research International Pilot Grant (P30 AI036214; Sub-award No. 112605914), the Canadian Institutes of Health Research (grants PJT-148621 and PJT-159625), and the Mexican Government (Programa Presupuestal P016; Anexo 13 del Decreto del Presupuesto de Egresos de la Federación).