HIV & cannabis — what the evidence actually says.

Cannabis is one of the few subjects in HIV care where the community got there decades before the research did. Long before the first randomized trial, people with AIDS used cannabis to eat, to sleep, to stop vomiting, and to get through nerve pain nothing else touched. That history is the reason medical cannabis programs exist in the United States at all.

Since then the science has caught up, and it tells a more textured story than either side of the old argument wanted: replicated randomized evidence for HIV-associated nerve pain, FDA-approved synthetic THC for appetite and weight loss, limited but reassuring data on antiretroviral drug levels, and real risks — especially around smoking, cardiovascular health, memory, and dependence.

The history — wasting syndrome, buyers' clubs, and Prop 215

Why HIV appears on nearly every state's medical cannabis list goes back to the late 1980s and early 1990s. Before effective combination antiretroviral therapy, AIDS wasting syndrome — involuntary weight loss, often with chronic diarrhea and fever — was among the most visible and feared features of advanced HIV disease. People lost twenty, forty, sixty pounds. Nausea from early antiretrovirals made eating a daily battle, and losing weight was not cosmetic — it predicted dying.

In that context cannabis was not a lifestyle choice; it was one of the only things that reliably made food possible. Physician Donald Abrams, who went on to run the first federally sanctioned trials of cannabis in people living with HIV, described exactly this: widespread smoked-cannabis use across the San Francisco Bay Area for HIV-related anorexia, weight loss, and treatment-related nausea was what prompted researchers to design a trial at all.1 The community practice came first. The trial protocol was written to catch up with it.

Cannabis buyers' clubs — informal, openly operated, membership-based dispensaries — sprang up in San Francisco and beyond, largely built and staffed by gay men, lesbians, and AIDS activists caring for their own. Abrams recorded that through the mid-1990s growing numbers of people with HIV were obtaining cannabis medicinally from local Cannabis Buyers' Clubs, and that federal attempts to shut that access down produced public outrage rather than retreat.1

Marinol arrives in 1992

The pharmaceutical response ran on a parallel track. Dronabinol — synthetic delta-9-THC in sesame oil, sold as Marinol — was already approved for chemotherapy-related nausea. In December 1992 the FDA extended approval to anorexia associated with weight loss in people with AIDS, and the label still reads that way today.2 Three decades later it remains the only FDA-approved cannabinoid medication with an HIV-specific indication.

For many people, a capsule of synthetic THC was not the same as the plant: oral dronabinol is slow, hard to titrate, and easy to overshoot, and if you are nauseated, waiting an hour is a poor substitute for something that works in minutes. That gap between the approved product and what people actually used is a big part of why the political fight happened.

Proposition 215, November 1996

In November 1996 California voters approved Proposition 215, making it the first U.S. state to endorse the medical use of marijuana. Abrams's contemporaneous account is blunt about the sequence: buyers' clubs, federal opposition, public outrage, organized medicine demanding research, then a National Institute on Drug Abuse–supported trial of the potential interaction between THC and the newly arrived protease inhibitors.1 Every state medical cannabis program that followed traces back through that door — opened primarily by people living with HIV and the people who loved them.

What cannabis is — and the Schedule I paradox

Some vocabulary, because the distinctions matter clinically and legally. Cannabis means products derived from Cannabis sativa, a plant with hundreds of chemical constituents including at least 125 cannabinoids.4 Marijuana, in U.S. regulatory usage, means plant parts or products with substantial THC; plants with very little THC are treated federally as industrial hemp.3

Delta-9-tetrahydrocannabinol (THC) is the most abundant intoxicating cannabinoid, responsible for most of the psychoactive effect — altered mood, thought, and perception.4 Cannabidiol (CBD) is non-intoxicating and widely sold in oils, supplements, foods, and lotions; it can still cause drowsiness, reduced appetite, vomiting, and diarrhea.4

What the FDA has and has not approved

The FDA has not approved the cannabis plant, or any product containing whole cannabis plant material, for any medical use.3 What it has approved are specific cannabinoid drugs:

• Dronabinol (Marinol, Syndros) — synthetic THC, approved for chemotherapy-related nausea and vomiting and for anorexia with weight loss in people with AIDS.2
• Nabilone (Cesamet) — a synthetic THC-like compound, approved for chemotherapy-related nausea and vomiting.3
• Plant-derived cannabidiol (Epidiolex) — approved for seizures associated with rare forms of epilepsy.4

The FDA has also determined that THC- or CBD-containing products cannot legally be sold as dietary supplements.3 That is why the gummy at the gas station is not a regulated medicine, and why an unlicensed product's label may not match what is in the package.

The scheduling paradox — and what changed in 2026

For more than fifty years the United States held two contradictory positions at once: dozens of states ran medical cannabis programs while federal law classified marijuana in Schedule I of the Controlled Substances Act — the category for substances with high abuse potential and no accepted medical use. That paradox is what made research so hard and left people nervous about asking.

In 2026 the federal picture shifted — partially. After a December 2025 executive order directing the Attorney General to expedite rescheduling, the Department of Justice and DEA published a final rule placing FDA-approved marijuana drug products and marijuana products regulated under a qualifying state medical marijuana license into Schedule III, alongside a new notice of hearing on transferring marijuana as a whole from Schedule I to Schedule III.5 DEA held that evidentiary hearing June 29 through July 15, 2026.5

ART interactions — the CYP pathways, in plain language

The enzymes involved

THC is metabolized by cytochrome P450 3A4 (CYP3A4) and also by CYP2C9. CBD is metabolized by CYP3A4. CBD in turn inhibits CYP2C19, and CYP3A4/5 inhibition has also been reported with CBD. Separately, smoking cannabis induces CYP1A2, increasing clearance of drugs metabolized by that enzyme — smoked cannabis increased clearance of theophylline by about 40% in one study, with the effect seen in regular use (more than two cannabis cigarettes per week) rather than occasional use.6

That matters in HIV care because CYP3A4 is also the main pathway for several antiretroviral classes and for the boosters ritonavir and cobicistat. When two drugs share a pathway, levels of one or both can move.

What was actually measured in people living with HIV

Theory says an interaction is plausible; measurement is better, and there is measurement — older, but directly on point. In a randomized study published in AIDS in 2002, people living with HIV on stable indinavir 800 mg every eight hours or nelfinavir 750 mg three times daily were randomized to smoked cannabis (3.95% THC), dronabinol 2.5 mg, or placebo, three times daily. Indinavir maximum concentration fell significantly in the smoked cannabis arm, but the changes for both protease inhibitors were small enough that the investigators concluded they were unlikely to carry short-term clinical consequences, and that neither cannabis nor dronabinol would be expected to affect antiretroviral therapy.7

The companion 21-day randomized, placebo-controlled trial in Annals of Internal Medicine in 2003 reached the same conclusion clinically: over three weeks, smoked and oral cannabinoids "did not seem to be unsafe" with respect to HIV RNA, CD4 and CD8 counts, or protease inhibitor levels.8

A Canadian HIV Trials Network pilot trial (CTN PT028) gave people living with HIV who had been virally suppressed for at least three years oral THC:CBD capsules for 12 weeks. The investigators did not exclude participants based on antiretroviral regimen, citing clinical experience in which no clinically significant interactions had been seen with heavy cannabis use — while still reviewing every participant's full medication list at screening.9 That is a reasonable model for how to think about it: not alarm, but attention.

Appetite and weight — Marinol versus the whole plant

Appetite is the oldest and best-established cannabinoid indication in HIV. Cannabis stimulates appetite through CB1 receptor activity, and was used for exactly that in AIDS wasting syndrome.9 Federal reviewers agree this is one area where cannabinoid drugs genuinely help: drugs containing cannabinoids may be helpful for loss of appetite and weight loss associated with HIV/AIDS, along with certain rare epilepsies and chemotherapy-related nausea and vomiting.3

Dronabinol — the approved option

Dronabinol is a capsule of synthetic delta-9-THC, FDA-approved for anorexia with weight loss in people with AIDS.2 The advantages are real: it is a prescription, often covered by insurance or an AIDS Drug Assistance Program, dosed to the milligram, smoke-free, and compatible with workplace policies that permit prescribed medication. So are the drawbacks: oral onset is slow and variable, effects can outlast what you wanted, and the intoxication can be harder to titrate than with inhaled products.

Whole-plant cannabis — what the trials showed

The 21-day randomized trial also looked at the immune system and found neither CD4 nor CD8 counts adversely affected; CD4 count actually rose significantly from baseline versus placebo in the smoked cannabis group, with a similar trend for dronabinol.9 Increased appetite is the most commonly reported physical effect of cannabis generally.4

What has never been done is a large, modern head-to-head trial of dispensary-grade whole-plant cannabis against dronabinol for weight gain in people living with HIV on effective treatment. And unintentional weight loss today is a reason to look for a cause — malignancy, opportunistic infection, depression, food insecurity, thyroid disease, medication side effects. Appetite support belongs alongside a workup, not instead of one.

Neuropathy — where the evidence is genuinely strongest

If there is one HIV-related symptom where cannabis has real randomized-trial support, it is painful HIV-associated distal sensory polyneuropathy — the burning, tingling, "walking on glass" pain in the feet and hands common among people who have lived with HIV for years, especially those exposed to older antiretrovirals.

Two independent randomized trials, two research groups, consistent direction, clinically meaningful responder rates in people whose pain had already failed other drug classes — a stronger evidence base than exists for much of what gets prescribed for neuropathy. It is also honest to name what these trials were not: small (fewer than 100 participants combined), short, conducted with smoked cannabis under direct observation, and using potencies far below much of today's market. The federal review of the wider literature describes the evidence for chronic pain overall as suggesting modest benefit.3

Pain, nausea, sleep, and anxiety — what people actually report

Beyond neuropathy, most of what we know about cannabis and day-to-day HIV symptoms comes from surveys and observational studies rather than trials. That is a real limitation and still informative, because it shows what people are treating and how well they think it works.

In a survey of 104 people living with HIV who use cannabis, conducted at a chronic viral illness service and published in Cannabis and Cannabinoid Research, median age was 54 and 88% were virally suppressed. Use ranged from more than once daily (32%) and daily (25%) to weekly (22%), monthly (17%), and rarely (6%). The most common reasons were pleasure (68%), pain (57%), anxiety (57%), and stress (55%). Forty-five percent rated cannabis "quite" or "extremely" effective for symptom relief. Secondary effects included feeling high (74%), increased cough (45%), paranoia (22%), palpitations (20%), and increased anxiety (21%) — and more than two-thirds said those effects did not bother them at all.11

Three things there deserve attention rather than judgment. Pleasure and symptom relief overlap heavily, so the sharp line between "medical" and "recreational" that state law depends on does not describe how people actually live. Cough at 45% is a smoking signal, not a cannabis signal. And roughly one in five reported increased anxiety and one in five palpitations — cannabis is not uniformly calming, and if it worsens your anxiety, that is pharmacology, not a personal failing.

The federal evidence review is measured on broader claims: modest benefit for chronic pain and multiple sclerosis symptoms, clear benefit for chemotherapy-related nausea and vomiting, and insufficient evidence for many other proposed uses.3 Sleep is among the most common reasons people use cannabis and one of the least studied outcomes. Anxiety cuts both ways: cannabis can cause anxiety, fear, distrust, panic, or hallucinations, especially with a large amount, a high-THC product, or little prior experience.4

Cannabis and viral suppression — does it get in the way?

This question carries the most fear, so separate the two mechanisms by which cannabis could affect viral load: pharmacology (does it change antiretroviral drug levels?) and behavior (does it change whether you take your medication?).

Pharmacology: no signal in the studies that exist

The 2003 randomized trial was designed to answer this, noting that cannabinoids could alter HIV RNA either by immune modulation or through cannabinoid–protease inhibitor interactions on shared cytochrome P450 pathways. Over 21 days, the adjusted average effect versus placebo on change in log₁₀ viral load was −0.07 for cannabis and −0.04 for dronabinol — statistically indistinguishable from placebo, with confidence intervals crossing zero — and neither CD4 nor CD8 counts appeared adversely affected.8 Protease inhibitor levels held up in the paired pharmacokinetic study.7

Some observational work points in an unexpected direction. A 2020 Clinical Infectious Diseases study found cannabis use among people living with HIV was associated with reduced systemic inflammation and faster decay of HIV DNA on antiretroviral therapy compared with people reporting no drug use, with no effect on cellular HIV RNA transcription.13 That is biologically plausible given cannabinoid effects on immune activation, but it is a cohort finding, not a treatment effect. Cannabis is not an HIV therapy.

Behavior: mixed, and it depends who you ask about

The adherence literature genuinely disagrees, and both sides deserve reporting. A Canadian cohort study in AIDS and Behavior followed 523 people living with HIV who used illicit drugs across 2,430 interviews between 2005 and 2012; 23.1% reported at least daily cannabis use at baseline. In both bivariate and multivariable analyses, at least daily use showed no association with optimal antiretroviral adherence (adjusted odds ratio 1.12; 95% CI 0.76–1.64; p=0.555). The authors concluded that cannabis may be used medicinally and recreationally without compromising effective adherence — while in the same analysis homelessness, daily alcohol use, daily heroin or cocaine injection, daily crack use, and incarceration were all negatively associated with adherence.14 The barriers that mattered were mostly structural.

A different picture emerged in an older population. A 2023 study in Open Forum Infectious Diseases found that among older people living with HIV, current cannabis users were 53% more likely to report less-than-perfect adherence, and among people who had missed a dose in the past week, 21% were current users versus 10% who had not used cannabis in the past year or ever.14

Smoked, vaporized, edible, sublingual — why the route changes everything

Route changes onset, duration, dose control, lung exposure, and which interactions apply. In the survey of people living with HIV who use cannabis, 97% smoked dry plant cannabis; 21% used edibles, 12% vaped, 12% used oils, and 2% used capsules.11 Smoking is still the default — and the route with the clearest documented respiratory harm.

Risks worth knowing — cardiovascular, cognitive, dependence, drug testing

Harm reduction means naming the actual harms, in proportion, without moralizing. Here is what the federal research summaries report.

Cardiovascular

Cannabis raises heart rate and blood pressure immediately after use, and some research links long-term use to increased risk of stroke, heart attack, and arrhythmias — though whether the connection is direct or driven by other factors is unsettled.4 That deserves weight in HIV care, because people living with HIV already carry elevated cardiovascular risk from chronic inflammation and often higher rates of tobacco use. If you have hypertension, coronary disease, or a statin on your list, say so out loud.

Cognition

Frequent or heavy use has been linked to problems with learning, memory, attention, processing speed, perceptual motor function, and language.4 That matters in HIV because neurocognitive complaints — the "brain fog" many people describe — can look similar. If your memory is slipping, cannabis is a variable a good workup should account for, not something to hide.

Mental health

Some evidence links cannabis to earlier onset of psychosis in people with genetic risk, and it may worsen symptoms in people who already have a psychotic disorder; high doses can induce a temporary psychotic episode. Some research shows increased depression risk among people who use cannabis during adolescence.4 None of that means a person with depression or anxiety cannot use cannabis — it means the interaction is worth tracking honestly.

Dependence potential

Chronic heavy use — daily or almost daily — of THC-containing products is associated with cannabis use disorder. Studies estimate 22% to 30% of people who use cannabis have it, and the strongest predictor is how often a person uses. One study estimated 12.1% of frequent users experience withdrawal — irritability, anxiety, restlessness, decreased appetite, depression, insomnia, unsettling dreams, headaches, sweating, abdominal pain, tremor. Cognitive behavioral therapy, motivational enhancement therapy, and contingency management can help; no medication is FDA-approved for cannabis use disorder.4

Two other risks surprise people. Cannabinoid hyperemesis syndrome — cycles of nausea, vomiting, and abdominal pain after long-term heavy use — often requires medical attention and resolves only with complete cessation.4 Cannabis use has also been linked to higher likelihood of head, neck, and throat cancer, particularly among people who smoke it.4

Drug testing — the practical landmine

THC metabolites are fat-soluble and can show up in urine for weeks after regular use. A state medical card does not override a federal employer's policy, a probation condition, a pain-management contract, a transplant program's requirements, or — in many states — a private employer's drug-free workplace rule. Secondhand exposure is documented too: in some environments secondhand cannabis smoke can produce positive drug test results, and positive urine tests have been reported in children exposed at home.4

Florida — how the medical marijuana card process actually works

Florida voters approved Amendment 2 in November 2016, and the program is run by the Florida Department of Health's Office of Medical Marijuana Use (OMMU). HIV/AIDS is explicitly on Florida's qualifying-condition list, alongside cancer, epilepsy, glaucoma, PTSD, ALS, Crohn's, Parkinson's, and multiple sclerosis — plus terminal conditions and chronic nonmalignant pain originating from a qualifying condition.15 That last clause matters: HIV-associated neuropathy is chronic nonmalignant pain originating from a qualifying condition.

The four steps, per OMMU

1. Be diagnosed by a qualified physician. You must be a permanent or seasonal Florida resident and be diagnosed with a qualifying condition by a physician who has completed the state's required training. That physician decides whether medical marijuana is an appropriate treatment.15
2. Be entered into the Medical Marijuana Use Registry (MMUR). Your physician enters you. Give them a current email address — OMMU sends two emails, one with your username and one with a temporary password.15
3. Apply for your Registry Identification Card. Apply electronically through the registry or by mail, with the $75 registration fee. You need an active card to purchase and possess medical marijuana. Cards renew annually, renewals are due 45 days before expiration, and a lost or stolen card costs $15 to replace.15 Card details and application instructions are on the OMMU Registry Identification Cards page.
4. Fill your order at a licensed Medical Marijuana Treatment Center (MMTC). Once your card is approved, contact an MMTC to fill the order your physician placed. Only state-licensed MMTCs may sell medical marijuana, and they may deliver if there is none in your city.15

Practical notes for Floridians living with HIV

• You don't have to use your regular HIV provider. Many HIV clinicians are not registered as qualified physicians in Florida. Get certified elsewhere and keep your HIV care where it is — but tell your HIV provider what you're taking, since they manage your antiretroviral regimen.
• Smokable flower is available to adults in Florida, though the statute sets special conditions for certifying a smoking form under age 18: a terminal condition, a physician determination that smoking is the most effective route, and concurrence from a board-certified pediatrician.15
• Your card is a Florida card. It authorizes purchase from Florida MMTCs — do not assume reciprocity elsewhere, and never carry cannabis across state lines or through an airport.
• Public use is not permitted under the state program, nor is purchase outside a licensed MMTC.15
• Adult-use cannabis is not legal in Florida. The program is medical-only, so the card is not a convenience — it is the legal basis for possession.

For help finding HIV care and wraparound services in Florida, see our Florida HIV resources page.

Nationally — the state map and where federal reform stands in 2026

State law is where medical cannabis access actually lives, and it is uneven. Some states run comprehensive programs with broad qualifying-condition lists that include HIV/AIDS; others run low-THC or CBD-only programs that would not reach most of what a Florida or California dispensary sells; a few have no medical program at all. The Marijuana Policy Project keeps a state-by-state tracker of medical, decriminalization, and adult-use status at mpp.org/states — the most practical single place to check your own state.

Three questions decide whether a state program is useful to you:

1. Is HIV/AIDS a listed qualifying condition, or must you qualify through chronic pain? Many programs list HIV/AIDS directly; others route people in through chronic or intractable pain, which changes the documentation you need.
2. What product forms are allowed? Low-THC/CBD-only states may permit oils but not flower or high-THC products — ruling out the routes used in the neuropathy trials.
3. Is there reciprocity? Some states honor out-of-state cards; many do not. If you split the year between states — common for Florida seasonal residents — check both.

Federal status as of 2026

The federal position moved in 2026 for the first time in half a century, and the change is narrower than the headlines suggested. After a December 2025 executive order directing the Attorney General to expedite rescheduling of medical marijuana, the Department of Justice and DEA published a final rule placing FDA-approved marijuana drug products and marijuana products regulated under a qualifying state medical marijuana license into Schedule III. DOJ also withdrew the prior 2024 hearing proceedings and noticed a new hearing on the broader Schedule I–to–III transfer, held June 29 through July 15, 2026.5

In practice, for someone living with HIV: research should get easier, because Schedule III removes some of the hardest barriers to clinical trials — the single biggest gap in this article. It does not mean cannabis is federally legal, that your state card travels, that federal employment or housing rules changed, or that interaction studies in modern high-potency products now exist. They don't yet.

Talking to your HIV provider — how to actually have the conversation

The most common reason people don't tell their HIV provider about cannabis is fear: of judgment, of a note in the chart, of losing access to pain medication, of a lecture. Those fears are not irrational. But your provider is the only person who can check your full medication list against CYP3A4, CYP2C9, and CYP2C19 and tell you whether anything on it should give you pause.6

Clinicians are increasingly expected to have this conversation. The Cannabis and Cannabinoid Research survey found the most commonly used information sources among people living with HIV were friends (77%) and the internet (55%) — and concluded that amid widespread regulatory change, health care providers should be prepared to answer questions about cannabis.11 If your provider isn't prepared, that is a gap in their training, not a verdict on you.

A script you can borrow

Plain, factual, no apology needed:

"I want to add something to my chart. I use cannabis — about [how often], mostly [smoked / vaped / edibles / tincture], mainly for [pain / appetite / sleep / nausea / anxiety]. I want to make sure it isn't interacting with my HIV meds or anything else I'm on. Can we look at my list?"

What to bring

• Route and frequency. Daily smoked flower and an occasional edible are pharmacologically different.6
• Product details if you have them. THC and CBD content, in milligrams if the label gives it. Knowing your cannabinoid content is associated with fewer negative cannabis-related consequences.11
• What you're treating. Naming the symptom lets your provider check for a better-studied option — including dronabinol, which has an FDA-approved HIV indication and may be covered.2
• Your full medication list — over-the-counter products, cotrimoxazole, SSRIs, warfarin, olanzapine or clozapine, transplant medication — the named interaction risks.6
• Side effects you've noticed. Palpitations, increased anxiety, worsening cough, memory problems, or cycles of vomiting all deserve reporting.411

If the conversation goes badly

You are allowed to change providers. A clinician who meets honest disclosure with stigma will not be a good partner in the rest of your HIV care either. Ryan White clinics, community health centers, and LGBTQ+ health centers tend to be most experienced at non-judgmental substance conversations. Community publications like POZ and TheBody have covered cannabis in HIV from a community perspective for decades if you want to hear from other people living with HIV first.