No history of drug overdose or recent illness was obtained. Upon arrival to the ED, the patient was obtunded , but would occasionally follow commands. Her Glasgow Coma Score was eight, scoring two points for eye-opening response, two points for verbal response, and four points for motor response. Vital signs revealed blood pressure of 135/81 millimeters of mercury , pulse 124 beats per minute , rectal temperature of 99.6 degrees Fahrenheit , and 97% pulse oximetry on room air. Physical exam revealed dilated pupils of six millimeters , normal neck exam, normal lung sounds, a soft and non-tender abdomen, and normal heart sounds. A neurological exam revealed rigidity in both lower extremities with a sparing of rigidity in the arms. Deep tendon reflexes showed sustained clonus in both feet, and the presence of hyper-reflexivity in the patella tendons bilaterally but with normal reflexes in the upper extremities. Lab results showed a normal complete blood count, normal creatine kinase, normal comprehensive metabolic profile, normal arterial blood gas, normal prolactin level,and a urine drug screen positive for THC. Electrocardiogram showed sinus tachycardia, and a non-contrasted head computed tomography was normal. Serotonin syndrome was considered in the differential diagnosis. After pediatric critical care and pediatric neurology consultation, one oral dose of cyprohepatidine 4 mg was administered. The patient was admitted to the pediatric intensive care unit. Magnetic resonance imaging of the brain was normal, and an electroencephalogram showed no epileptic activity. The patient rapidly improved and was discharged the following day. Prior to discharge, the patient admitted to “dabbing” about 30 minutes prior to arrival to the hospital. The same patient returned to the ED the following night with a similar presentation, once again associated with dabbing. Over time, improvements in antiretroviral therapy have lengthened lifespan and reduced HIV transmission among people living with HIV. Findings that ART adherence can suppress viral load and reduce HIV transmissibility during condomless sex have led to prioritizing treatment as prevention as a key strategy to prevent HIV transmission by PLH.

However, high rates of substance use and depression among PLH remain key barriers to successful implementation of TasP in the U.S. and other similar settings. However, empirical studies on the associations of substance use,indoor plant table depression, and achieving undetectable VL have not been adequately assessed in low- and middle-income settings. Non-injection substance use is the most common form of substance use among PLH, with 40–70% reporting the use of alcohol, cannabis, non-injection stimulants , and/or opioids. In general, PLH who use substances are less likely to access ART, are found to have lower ART adherence, are less likely to achieve viral suppression, and are more likely to have faster disease progression compared to non-substance using PLH. Moreover, this population may be the most likely to engage in condomless sex, making it critical to understand how to improve their HIV care outcomes. Aside from behavioral risk, and the reduced ART adherence associated with substance use, emerging research indicates that substance use may have pathophysiological effects on HIV disease progression. For example, stimulants have been linked to increased HIV replication—in peripheral blood mononuclear cells and in mouse models. When examining the effects of substance use on VL or other HIV outcomes, it is also important to investigate the contribution of depression as it is a highly prevalent comorbid condition. Depression is a more common comorbidity to substance use among PLH than the general population, and is the most common psychiatric health condition among PLH—affecting 20–33% of adults in HIV care. In terms of HIV clinical outcomes, depression is thought to lower ART adherence and reduce the likelihood of sustained viral suppression. Studies indicate that depressive symptoms may also affect HIV disease progression above and beyond sub-optimal ART adherence by reducing individuals’ responsiveness to ART, decreasing CD4+count, and increasing HIV VL. Depressive symptoms and substance use are prevalent among PLH and likely contribute substantially to the lack of sustained viral suppression. Despite the high prevalence of substance use and depressive symptoms among PLH, most research examining depression, substance use, and HIV disease outcomes has been conducted in the U.S..

There is little information on the type and patterns of non-injection substance use, on the prevalence of depression, and on how these common comorbidities affect viral suppression among PLH in low- and middle-income settings. There is reason to think that the association between substance use and viral load detectability may operate through decreased ART adherence and increased co-morbidity with depression among PLH. Previous research have linked substance use—including alcohol, cocaine, heroin, methamphetamines, and other stimulants—to decreased ART adherence, although these studies took place in the U.S.. A recent systematic review focused on ART adherence among those who engaged in substance use in low- and middle income countries found sub-optimal adherence to treatment, however this review solely focused on injection drug use. In addition, a study that examined active drug use on ART adherence and viral suppression found that depression appeared to mediate the association, although the finding was only significant for HIV-infected women and not HIV infected men. Moreover, based on the minority stress theory—which posits that sexual minorities have adverse health outcomes as a result of heightened stress from prejudice and stigma based on their sexual minority status—it is thought that men who have sex with men may have greater substance use and depressive symptoms than heterosexual men. This greater comorbidity prevalence is hypothesized to magnify the association between substance use, depression, and viral load detectability. This is likely the case for men in low- and middle-income settings, such as Thailand and Brazil, where HIV prevalence is much greater among MSM compared to the general adult population at 9.2% and 10.5% , respectively. Although less research has been conducted among men who identify as heterosexual in international contexts, in Brazil they comprise the largest proportion of men infected with HIV and as many as 70% receive late HIV-diagnosis. Furthermore, non-injection substance use often affects MSM and heterosexual men at greater rates than women, potentially exacerbating the effects of substance use on HIV outcomes via ART adherence and depression in low- and middle-income settings.

This study aims to address this gap in research by conducting a secondary data analysis focused on MSM and heterosexual men using HPTN 063 data, a longitudinal observational study of HIV-positive individuals in HIV care in Zambia, Thailand, and Brazil. First, we described the type and pattern of non-injection substance use and prevalence of depressive symptoms among men infected with HIV at baseline. Second, we examined the effect of non-injection substance use on ART adherence and HIV VL undetectability, testing ART adherence as a mediator of the association between substance use and HIV VL undetectability. Third, we examined the effect of non-injection substance use on depressive symptoms and VL undetectability, testing depressive symptoms as a mediator of the association between substance use and HIV VL undetectability. Then, we tested whether there was evidence of effect modification due to sexual orientation, on the association between substance use, mediators , and HIV outcomes. For all analyses, we stratified by unique country context.Data were collected via HPTN 063, a multi-site, longitudinal observational cohort study of people living with HIV at high risk for sexual transmission in HIV care in Africa , Asia , and South America . Recruited participants included HIV-infected heterosexual men, heterosexual women, and men who have sex with men . Structured interviews were conducted every 3 months over the course of 12 months, collecting data on socio-demographics, behavioral risk, substance use, mental health, and ARV adherence. HIV clinical variables were extracted from patient flews.The HPTN063 study design has been described in detail in previous publications.Plasma HIV–RNA VL was extracted from medical records at baseline and each follow-up visit and recorded if a current VL was documented. VL was then dichotomized . Non–injection substance use was measured as the number of self-reported use days and included stimulants, cannabis, and alcohol. Stimulant use was measured as the number of days that non-injection cocaine , methamphetamine, and ecstasy use were reported in the prior 3 months. Cannabis was measured as the number of days that marijuana and hashish were reported in the prior three months. Alcohol misuse was measured using the 10-item alcohol use disorders identification test . Example items include how many drinks containing alcohol one has on a typical day and how often one is not able to stop drinking once started. AUDIT score was dichotomized into alcohol misuse versus no alcohol misuse. Polysubstance use was measured as the total number of non-injection substances reported used in the past 3 months , including stimulants, cannabis, and alcohol misuse, plant growing stand and was treated as a continuous variable . Depression symptoms were measured using the Center for Epidemiologic Studies Depression Scale. Example items ask how often during the past week participants had a poor appetite or felt depressed. CESD score was dichotomized into severe depressive symptoms versus not severe depressive symptoms . ART adherence was measured using the self-reported question on adherence ability, “in the last 3 months, on average, how would you rate your ability to take all your antiretroviral drugs as your doctor prescribed?”.

Instructions provided prior to the interview normalized ART non-adherence. Participants were provided with a response card with Likert response options, ranging from very poor to excellent. This single-item, self-report adherence measure has been found as valid and reliable in prior research. Due to small cell size, ART adherence ability in Thailand was recoded into two levels . For Brazil, ART adherence ability was missing on too many participants to warrant inclusion in this analysis and the dichotomized variable of taking ARTs was used in place. The self-reported measure asked, “In the last 3 months, have you taken antiretroviral drugs?” Socio–demographic variables included in our analysis were age group and education .Data analysis began with descriptive statistics at baseline of the total sample and of heterosexual men versus men who have sex with men on non-injection substance use, depression, HIV outcomes, and socio-demographics. The Chi square statistic test was used for categorical variables, and t-statistic test for continuous variables, to detect statistically significant differences between groups . Next, we described the type and number of self-reported non-injection substances used in the prior 3 months at baseline stratified by country and sub-group to understand poly-substance use in our sample . Then, generalized linear mixed models were applied with the logit link function for longitudinal binary outcomes to estimate the odds ratios of non-injection substance use on having an undetectable HIV VL adjusting for covariates, age and education . The mediators, ART adherence and depression, were also estimated as an outcome of non-injection substance use using GLMM and mediation was controlled for when estimating the effects of non-injection substance use on undetectable HIV VL. GLMMs with the logistic link function with a random intercept and compound-symmetric covariance were used to account the correlations of observations between visits within individuals. All analyses were stratified by country. For each model, an interaction term of substance use and sub-group was included to test for statistically significant differences between MSM and heterosexual men in the associations between substance use and ART adherence, depressive symptoms, and undetectable VL.Table 1 shows the baseline characteristics of participants in the total sample stratified by study site and heterosexual men versus MSM. In Thailand, 43% of the total sample reported alcohol misuse. In the past 3 months, individuals, on average, reported using stimulants for zero days , cannabis one day , and used one non-injection substance , with no significant difference by sub-group. Twenty-two percent of the total sample had severe depressive symptoms, with no significant difference by sub-group. In terms of HIV outcomes , 82.4% reported good/very good/excellent adherence ability, with MSM reporting significantly better adherence ability than heterosexual men . Seventy-seven percent of the total sample presented an undetectable VL at baseline, with no significant differences by sub-group. The median CD4+ count at baseline was significantly lower among heterosexual men compared to MSM . In Brazil, 34% of the total sample reported alcohol misuse. In the past 3 months, individuals, on average, reported using stimulants for 4 days , cannabis for 5 days , and used one non-injection substance , with no significant difference by sub-group. About half of the sample in Brazil had severe depressive symptoms, with no significant difference by sub-group.