The models for the HIV+ group examined HIV-related clinical factors including: CD4+ cell count, detectable HIV viral load status and HAART use. In addition, for the analysis in the men in the early cohort, we estimated models for the period 2002–2013 in order to better compare the results with the men in the late cohort. Missing data for correlates were imputed using multiple imputation with chained equations . Five imputed datasets were generated for missing baseline and time-varying correlates which range from 0.2% to 14.8% and the estimates were combined according to Rubin’s rules . Because of the large sample size and number of person-visits, small prevalence ratios may be statistically significant. Thus we calculated a measure of effect size for the adjusted prevalence ratios . Cohen h or d of 0.2, 0.5 and 0.8 are small, medium and large effect sizes respectively . Throughout the analyses, P values were not adjusted for multiple comparisons. However, we highlight results where effect sizes equal or exceed the criteria for ‘small’ effects . Statistical analyses were performed using SAS version 9.4 and STATA version 11.Table 1 displays the baseline characteristics of the 5,914 participants in this study stratified by HIV serostatus and cohort enrollment. The mean age at baseline ranged from 33 years [standard deviation =6.7] among the HIV+ men in the early-cohort to 39 years among the HIV+ men in the late-cohort. The men in the early-cohort were predominantly non-Hispanic, white , cannabis vertical farming whereas the majority of the men in the late-cohort were non-Hispanic, black . At baseline, the men in the early-cohort were more educated than the men in the late-cohort .
At baseline, the prevalence of marijuana use was highest among the HIV+ men in the early-cohort and lowest among the HIV+ men in the late-cohort . However, daily marijuana use, among current users, was highest among the HIV+ men in the latecohort and lowest among the HIV− men in the early-cohort .Among the men in the early cohort, the annual prevalence of current marijuana use declined significantly from 80% in 1984 to 33% in 2013 among the HIV+ men and from 58% in 1984to 22% in 2013 among the HIV− men . The prevalence of daily marijuana use among current users increased significantly from 14% in 1984 to 32% in 2013 among the HIV+ men and from 9% in 1984 to 22% among the HIV− men . Among the men in the late cohort, prevalence of current marijuana use declined modestly from 32% in 2002 to 29% in 2013 among the HIV+ men, and decreased significantly from 37% in 2003 to 26% in 2013 among the HIV− men . However, daily marijuana use among current users increased significantly from 17% in 2002 to 37% in 2013 among the HIV+ men and from 16% in 2002 to 34% in 2013 among the HIV− men . Overall, the prevalence of daily marijuana use among all men in both the early and late cohorts were relatively stable across the followup period . The number of observations contributing to the yearly prevalence estimates for each cohort are displayed in supplemental tables S3 and S4. In this analyses of the MACS cohort, the annual prevalence of current marijuana use decreased over time among all men . However, in contrast, daily marijuana use, among those who used marijuana in the previous six months, increased among the HIV+ and HIV− men in both the early- and late- cohort enrollment: increasing by more than two-folds in nearly all groups.
Among the participants enrolled before 2001 in the MACS, the HIV+ men reported significantly higher prevalence of current and daily marijuana use as compared to the HIV− men with results reaching Cohen’s small effect size but no significant difference in marijuana use by HIV serostatus among the men enrolled after 2001. Alcohol use, particularly heavy alcohol use was significantly associated with current marijuana use and reaching Cohen’s small effect size in the analyses for both the early- and late- cohorts. Completing a graduate work or more was negatively associated with daily marijuana use and reaching Cohen’s small effect size in the analyses for both the early- and late- cohorts. The prevalence of marijuana use increased after passage of a MML in the analysis that included all men in the early-cohort but not for the men in the late-cohort, though these results did not reach Cohen’s small effect size. None of the significant associations between HAART use, CD4+ cell count and detectable HIV viral load and prevalence of marijuana use reached Cohen’s small effect size. The contrasting decline in annual prevalence of current marijuana use but increasing prevalence of daily marijuana use among users found in the current study is consistent with recent data from HIV+ women in the Women’s Interagency Study , where the authors found that between 1994 to 2010, there was a significant decrease in prevalence of current marijuana use from 21% to 14%. The most plausible explanation for the declining trend in current marijuana use may be the advancing age of participants in the MACS. One likely explanation for the increase in daily use may be that occasional marijuana users declined use over time.
The relatively stable trend in the prevalence of daily use over time supports this explanation. Between 1984 and 2013 – the period of this study – 3 of the 4 states that have MACS sites passed laws legalizing marijuana for medical purposes. In recent years, attitudes about marijuana use in the US have tempered and there has been an increase in population acceptance of marijuana use . Though not reaching Cohen’s small effect size, among the men in the early-cohort, passage of a MML was associated with an increase in the prevalence of current marijuana use in the analysis including all men, but not in the analysis that included only the HIV+ men. It is possible that the HIV+ men in the early-cohort may have already formed attitudes regarding marijuana use that passage of a MML did not influence their use. Among the men in the late-cohort, passage of MML was not significantly associated with increased marijuana use. This finding may in part be due to the short time periods pre and post enactment of the laws which may not have provided sufficient time to detect a change in their prevalence of use. Among the HIV+ men in current study,grow cannabis in containers there were few significant associations between HAART use, CD4+ cell count, detectable viral load and prevalence of marijuana use and of those that were significant none reached Cohen’s small effect size. These findings are similar to prior studies that report no significant or clinically meaningful differences in HIV viral load or CD4+ cell count among marijuana users as compared to nonusers. Yet others have found significantly lower HIV viral load and higher CD4+ count in marijuana users, although these studies differ methodologically as well as in the samples included. Taken together, these findings underscore the complex relationship between marijuana use and markers of HIV disease stage/progression and therefore warrant further study. There are some limitations to our study. We relied on self-report of marijuana use and no biological marker of marijuana use was used to confirm the self-reported data. Furthermore, data in the current analyses was collected from an ongoing longitudinal study with extended follow-up, thus attrition due to death or loss to follow-up may have influenced the prevalence estimates. This study did not assess prevalence or trends in marijuana use disorder, recreational versus medical use or other parameters of marijuana use . Also, it is important to note that the effects for passage of MML and marijuana use reported in this study should not be interpreted as causal. Our study included only four states and two states had insufficient time windows pre and post enactment of laws to provided enough information to discern a change in trend. Despite these limitations, our study has notable strengths. Our study utilized data from a large and diverse sample of HIV+ and HIV− MSM with extensive follow-up period to assess changes in prevalence of and correlates for marijuana use. In the U.S., use of prescription pain relievers , also known as prescription opioids and opioid pain relievers, has been increasing dramatically.
Worldwide, prescriptions of PPRs have almost tripled since 1990, and the U.S. is a factor in this rise, as it has the highest per capita consumption of PPRs in the past ten years . This increase has become dangerous, as opioid use carries risks that include addiction, sedation, respiratory depression, overdose and death . Between 1999 and 2010, deaths attributed to PPRs rose five times among women and 3.5 times among men . Of all prescription drug OD deaths in the U.S. in 2013, 71.3% involved PPRs . PPRs and marijuana are biologically linked; like PPRs, marijuana induces analgesia, acts on some of the same brain regions, and partly exerts its effects via opioid receptors . This connection is especially relevant due to the changing legal status of marijuana. As of August 2016, 24 states and Washington D.C. had legalized medical marijuana. Between 2007 and 2012, the number of past month marijuana users rose from 5.8 to 7.3% 2013), and between 2001 and 2013, past year adult marijuana use increased from 4.1 to 9.5% in the U.S. . Further, legalization of medical marijuana has been associated with increased odds of marijuana use among adults , though no consistent association has been determined among youth/young adults . Distinct theories attempt to explain how medical marijuana legalization affects use of substances other than marijuana. The relationship between different substances can be impacted by 1) change in cost of a substance, 2) policy alterations that influence availability of a substance, 3) shifts in legal consequences of using a substance, and/or 4) the psychoactive/pharmacological effects of a substance . More U.S. states are legalizing medical marijuana , and marijuana shares some psychoactive/pharmacological effects with PPRs. The substitution theory postulates that there is a substitution effect, whereby an increase in marijuana use coincides with a decrease in the use of other substances – in this case, PPRs . There are logical reasons why individuals would opt to use marijuana instead of PPRs. With the new legal status of medical marijuana, individuals can access it through medical dispensaries and enjoy a lower legal risk if they live in a state where it is legalized. Individuals also report switching to marijuana for pain control because when compared to prescription drugs, marijuana has fewer side effects and withdrawal symptoms . Studies supporting the substitution effect have demonstrated that either increases in the use of marijuana or the legalization of medical marijuana is associated with reductions in opioid use, hospitalizations for opioid dependence/abuse, PPR ODs, and opioid OD mortality . In contrast to the substitution effect, there may be a complementary effect, where an increase in marijuana use is associated with an increase in the use of PPRs . In support of this theory, researchers using National Survey on Drug Use and Health data found a positive association between marijuana and increased use ofPPRs . In another study, researchers focused on individuals who were prescribed long-term opioid therapy and found that those who also used medical marijuana presented with greater risk of misusing prescription opioids. Additionally, a prospective cohort study using the National Epidemiologic Survey of Alcohol and Related Conditions data determined that use of marijuana was associated with a greater risk of using nonmedical prescription opioids three years later . However, in these studies, researchers did not analyze how co-use of other substances would impact the direction and/or strength of the relationship between marijuana and opioids/PPRs. To determine if there is either a substitution or a complementary effect between marijuana use and PPR use, co-use with other substances needs to be studied. Additionally, there is a strong positive association between nicotine use and PPR use. When compared to non-smokers, tobacco smokers experience more intense and longer lasting chronic pain, as well as a higher frequency of PPR use . Studies have demonstrated an interaction between nicotine and opioids that is associated with an increase in the total consumption of the two substances and contributes to other effects of the drugs .