The present research revealed the prevalence of illicit drug use as well as the particular patterns of consumption among PLHIV in Spain. The most frequently consumed substances among PLHIV in Spain, according to our results, were: alcohol (86.7%), tobacco (55.0%), illicit drugs (49.5%) and other substances (27.1%). The prevalence of last-year alcohol consumption among PLHIV was higher than that reported for the general Spanish population (77.6%), although daily consumption resulted in similar results [35]. The prevalence of active smoking was high compared with the general population (30.8%) [35]. Studies in other countries focusing on PLHIV have reported similar results to those we found [36,37].

Almost half of the participants in our study had used illicit drugs in the last 12 months, with cannabis, cocaine, and poppers being the most frequent ones. Cannabis and cocaine are also the most consumed drugs in the general population [35,38] but the prevalence found in our results was three times higher for cannabis and 13 times higher for cocaine than the respective figures reported by the European Monitoring Centre for Drugs and Drug Addiction (EMCDDA) in the general population [10]. As for studies focusing on PLHIV, the prevalence found in our study was slightly higher compared with the results by Garin et al. [9]. In addition, the prevalence of cocaine use among MSM who consume drugs in our sample was higher than that found recently by González-Baeza et al. [7], but results were similar for most of the other substances, such as poppers or ketamine. At an international level, several studies have also highlighted the significant prevalence of illicit drug use in PLHIV across Europe. For example, our results were similar to those found by Daskalopoulou et al. [5] in the United Kingdom, although direct comparison is not possible since that study evaluated last-three-months consumption. Grabovak et al. [39] found a moderately higher prevalence of drug use in PLHIV in Austria but their analysis included sildenafil/tadalafil, categorized as ‘other substances’ in our study. In contrast, our prevalence results were higher compared with other studies in France [40] or Scotland [37] although, again, differences in time frame make comparisons.

Our study also found some differences in the pattern of drug use between the two largest cities in Spain (Madrid and Barcelona). Similar results were found in a study conducted in 2016 that collected data from 486 MSM in Spain, who had practised ‘chemsex’ in the previous 12 months [41]. Findings of the EMIS study showed that the prevalence of use of the different drugs varies substantially across cities, suggesting that drug use patterns appear to be culturally and socially determined [42]. The association of the city of residence with drug use could be influenced not only by psychological factors, but also by structural drivers such differences in drug availability and differences in national, regional, or local norms and cultures concerning drug use [42,43].

Regarding routes of consumption, we found that near to 7% of PLHIV who used illicit drugs were using the injection route. They represented 3.4% of the total sample. This percentage is comparable to the HIV incidence results reported in Spain and the countries from the European Union and European Economic Area [1,44].

Regarding patterns of drug use, our results found four clusters of PLHIV who used drugs. Two of them were composed mostly of HTX and the other two principally of MSM, which resulted in being the most prevalent group of PLHIV who used illicit drugs. MSM were the youngest participants, had the most recent HIV diagnosis and the best socio-economic status. These socio-demographic and epidemiologically different characteristics between MSM and HTX living with HIV are concordant with those existing in the Spanish national HIV data about HIV prevalence and incidence [1,2,45]. MSM and HTX clusters of PLHIV who used illicit drugs found in the present study also differ in their reasons for using drugs. While HTX aimed at avoiding worries or negative feelings, MSM pointed to issues such as enjoying their use and sexual purposes.

Regarding clusters of HTX who used illicit drugs, the largest one was composed of people who presented the lowest average of polydrug use and who were users of cannabis mainly followed by cocaine powder. More than a half of them were former PWID and nearly one-fifth of them were taking methadone in the period assessed by our study. Thus, they have given up or replaced the use of heroin (or other injected drugs). In addition, a small group of HTX was also found who were polydrug users and who showed the highest prevalence of heroin use and current injection route. Although the majority of PLHIV in this cluster acquired HIV infection through sharing needles, during the study period only near a third of them continued injecting drugs. Taking the previous results conjointly, we can conclude that according to previous literature [19], both the use of heroin and its intravenous use have continued to decrease in Spain. In our study, PLHIV in CL2 (mainly users of heroin and cocaine) represented around 6% of the total sample and among them, 2% were PWID. Despite this decrease, it is essential to pay attention to these groups of people because we found they presented the lowest levels of adherence to treatment [20,21]. With this regard, PLHIV who were users of heroin and cocaine presented the poorest biological markers of HIV infection. However, not only the impact of their use of drugs but also their socio-epidemiological profile could explain these results. The literature shows that PLHIV who use illicit drugs are more likely to experience problematic adherence and that they have increased age-matched morbidity compared to those who do not use drugs [46,47]. A systematic review of interventions to improve adherence and virological outcomes among PLHIV who used drugs concluded that directly administered antiretroviral therapy (DAAT) obtained the most substantial evidence of efficacy but that it is labour-intensive and costly to implement [46]. The arrival of long-acting treatments could be an opportunity for DAAT in PLHIV who use illicit drugs. Nevertheless, DAAT should be complemented with other interventions that have shown efficacy, such as cognitive behavioural therapy, motivational interviewing and peer-driven interventions [46]. Besides interventions to promote adherence, pharmacokinetic drug interactions must be adequately addressed in people who use illicit drugs [47].

Among the MSM using drugs, there was a majority that reported moderate consumption, both regarding frequency and diversity of drugs used. However, there was another pattern of use in MSM, including younger patients with the highest frequency of illicit drug use and relevant polydrug consumption. Some of them used drugs intra-rectally (around 14%) and intravenously (around 10%). PWID represented around 2% of the total MSM in our study. This percentage was lower than that found by González-Baeza et al. [7] (around 4%) and similar to the one found by Folch et al. [11] (1.4%). Prevalence of PWID in Folch et al. [11] was higher in the largest towns. Baeza et al. [7] conducted their study exclusively in MSM with HIV living in a large Spanish city (Madrid), where the use of recreational drugs and the practice of ‘slamsex’ seems to be higher than in smaller cities [11,18]. Regarding other countries, Pufall et al. [13] found that 10% of the MSM interviewed in UK practised slamming in the previous year. Although MSM clusters reported frequent use of drugs in a sexual context (32%) and high rates of STIs in the previous 12 months, the percentages were the highest in MSM polydrug users. These results are similar to those reported in the study of González-Baeza et al. [7]. Recent studies are highlighting the importance of the ‘chemsex’ phenomenon and its health consequences among those who practise it [13,48,49]. Thus it will be necessary to monitor the impact of drug use in the health and psycho-social sphere of MSM who use illicit drugs. It will also be necessary to implement coordinated and multi-dimensional strategies that reduce drug-related harm among them. Specific interventions could include: direct client service support to manage or reduce the use of drugs, health promotion activities, peer education support, training healthcare providers and drug services about new recreational drugs and their role in sexual intercourse, services to address specific psychological and psychiatric problems in MSM who practice ‘chemsex’, advice about healthy leisure alternatives, harm-reduction services such as needle exchange, among others. [41,50,51]. Moreover, several study findings suggest that MSM who combine drugs with sex are more likely to engage in high-risk sexual practices when compared with MSM who do not combine drugs with sex [43]. Thus, it will be necessary to examine and increase the use of bio-medical risk reduction interventions such as STI testing, post-exposure and pre-exposure prophylaxis to reduce the potential negative consequences of ‘chemsex’ on public health [43].

Our study has several limitations. First, its cross-sectional nature does not allow causal relationships to be established. Second, we did not use a probabilistic sampling, although the large sample size and extensive geographical data collection could mitigate this limitation. In this sense, the socio-demographic characteristics of our participants were concordant with Spanish epidemiological data for PLHIV [1]. Also, we aimed to provide a general overview of drug use, including different PLHIV subgroups in Spain. This may make difficult direct comparison with studies focusing on MSM only. However, the sample in our study involved a large proportion of MSM, similar to that found in other studies [7], and specific analysis and results are provided for this subgroup. Furthermore, we used a self-reported data collection method, and this could underestimate the prevalence of PLHIV using drugs because some people would not desire to reveal this information to their health-care providers. However, we tried to overcome this limitation through the anonymity of the survey and by balancing the data collection from hospitals and NGOs. Furthermore, another limitation was that we measured the use of substances using single items. However, it allowed us to assess the use of several substances. Moreover, other studies had concluded the utility of single screening questions to identify substance use [52]. Likewise, our survey asked about gender categories (men, women, transgender) considering them as mutually exclusive. However, many transgender people identify as men or women, not as a third gender.