HIV / AIDS · Chapter 1
The HIV Decision Framework
How to think about psychotropic prescribing when HIV is the constraint — why the drug interaction, not the dose, is the axis that matters; why most modern regimens are cleaner than their reputation; and the trap the other courses never had: a brain infection wearing a psychiatric mask.
Bottom Line Up Front
The 30-second version
- The axis changes from the other courses. Cardiac, renal, and hepatic disease bend the dose. HIV bends the drug interaction. The controlling question is not “how much do I reduce this?” but “what is the antiretroviral regimen doing to my drug — and what are the two doing together?”
- Rule out the infection before you reach for the prescription pad. New psychiatric symptoms in someone with a CD4 count under 200 are a CNS opportunistic infection until proven otherwise. Toxoplasmosis, cryptococcal meningitis, PML, and CMV all present with behavior change, psychosis, or confusion. Image and tap first; prescribe second.
- Three buckets tell you which way the levels move. Pharmacokinetic boosters (ritonavir, cobicistat) push psychotropic levels up; NNRTI inducers (efavirenz) push them down; and modern integrase-based regimens (bictegravir, dolutegravir) mostly leave them alone. Name the bucket and you have your direction.
- The 2026 regimen is cleaner than its reputation. First-line therapy is now integrase-inhibitor-based, with relatively few interactions — so the danger has migrated to legacy boosted and efavirenz regimens you inherit, and to a handful of specific collisions (carbamazepine and dolutegravir the sharpest). Reflexively fearing every antiretroviral is as wrong as ignoring them.
Which Axis to Trust (60-Second Version)
The three prior courses trained one reflex: find the number, reduce the dose. HIV asks a different question. Antiretrovirals rarely make you lower a psychotropic dose for clearance reasons — they change the drug's plasma level by inducing or inhibiting the enzymes that metabolize it, and they add their own pharmacodynamic load on top. So the point-of-care move is not to grade an organ; it's to read the regimen.
Two axes run in parallel, and you have to hold both. The pharmacokinetic axis asks what the antiretroviral does to the psychotropic's concentration: a booster inhibits CYP3A4 and levels climb; an NNRTI induces and levels fall. The pharmacodynamic axis asks what the two drugs do together regardless of level — additive QTc prolongation, additive sedation or CNS toxicity, additive serotonergic tone, a lowered seizure threshold. The pharmacokinetic number can look reassuring while the pharmacodynamic stack is what harms the patient.
Why not just “reduce the dose”?
Because the direction isn't fixed. Depending on the regimen, the same antidepressant may need to go up (an inducer is stripping it away) or down (a booster is stacking it). A blanket “go low in HIV” rule will underdose half your patients and overdose the other half. You have to know which enzyme story you're in — and that comes from the regimen, not a lab value.
Rule Out the Infection First (the Gate No Other Course Has)
Before any of the pharmacology matters, the single highest-yield move in HIV psychiatry is to confirm you are treating a psychiatric illness and not a brain infection. Opportunistic infections of the CNS present with exactly the symptoms that land on a psychiatry consult — new psychosis, personality change, confusion, mood shift — and they cluster by how far the CD4 count has fallen. As a rule, opportunistic infections begin once the CD4 count drops below 200 cells/µL, and the neuropsychiatric ones deepen from there.
| CD4 (cells/µL) | Consider | How it can look on a psych consult |
|---|---|---|
| < 100 | Toxoplasma encephalitis | Focal deficits with headache or seizure — but also bizarre behavior, hallucinations, and personality change. (The general OI gate is < 200; toxoplasmosis specifically clusters < 100.) |
| < 100 | Cryptococcal meningitis | Subacute headache and malaise, apathy, blunting, confusion — often mistaken for depression before the meningismus is obvious. Frequently presents at deeper counts (< 50). |
| < 200 | PML (JC virus) | Aphasia, hemiparesis, or ataxia progressing over days to weeks; cognitive and behavioral change, typically without fever. |
| < 50 | CMV encephalitis | Delirium and rapidly progressive confusion at the deepest immunosuppression. |
The practical posture: in anyone with a CD4 under 200 (or unknown) and new neuropsychiatric symptoms, the framework is neuroimaging and, where indicated, a lumbar puncture before you attribute the picture to a primary psychiatric disorder. HIV itself, immune reconstitution inflammatory syndrome, and antiretroviral neurotoxicity round out the differential. Prescribe an antipsychotic for “new psychosis” that is actually toxoplasmosis and you have treated the symptom and missed the disease.
The Three-Bucket ART Model That Drives the Decision
Child-Pugh gave the hepatic course a single tool to anchor every dose. In HIV, the anchoring tool is the metabolic personality of the regimen. Every antiretroviral a psychiatrist worries about falls into one of three buckets, and the bucket tells you which direction your psychotropic level will move.
| Bucket | Agents | Effect on psychotropic levels |
|---|---|---|
| Boosters / inhibitors levels UP | Ritonavir, cobicistat — the “/r” and “/c” in a regimen (darunavir/ritonavir, atazanavir/cobicistat, elvitegravir/cobicistat) | Potent CYP3A4 inhibition raises levels of many psychotropics. Classic contraindication: oral midazolam and triazolam. One counterintuitive exception — bupropion actually falls with ritonavir (which also induces CYP2B6), though not with cobicistat, which lacks that inducing effect. |
| Inducers levels DOWN | Efavirenz, etravirine, nevirapine (the NNRTIs) | Enzyme induction lowers psychotropic levels — risking loss of efficacy. Efavirenz also carries its own neuropsychiatric burden. |
| Clean anchors mostly no effect | Integrase inhibitors bictegravir and dolutegravir; the NRTI backbone (tenofovir, emtricitabine, lamivudine); rilpivirine, doravirine, maraviroc | Relatively few interactions — the modern first-line default. Watch the reverse direction: a psychotropic inducer (carbamazepine) can strip dolutegravir. |
The single most useful habit: find the booster or the NNRTI first. Scan the regimen for a “/r” or “/c” (booster — levels up) or for efavirenz/nevirapine/etravirine (inducer — levels down). If you find neither and the anchor is bictegravir or dolutegravir, you are probably in a low-interaction regimen — and your attention shifts to the pharmacodynamic axis and the one bidirectional trap.
The Core Principle: There Is No eGFR for HIV
This is the idea to carry out of the chapter. The kidney hands you a continuous filtration estimate; the liver hands you a categorical class. HIV hands you neither — there is no lab value that tells you how a given psychotropic will behave. The CD4 count measures immune damage and the viral load measures control, but neither predicts a drug interaction. The thing that governs your prescribing is the named regimen, drug by drug.
So the point-of-care question is not “How advanced is the HIV?” — it's “What exact antiretrovirals is this patient on, and what does each one do to enzymes?” Three things complicate the clean picture:
- The booster paradox. Ritonavir inhibits CYP3A4 — but it induces other pathways (UGT, CYP2B6), so a few substrates, bupropion the notable one, drop instead of rise. Cobicistat, the other booster, lacks that induction, so the paradox is ritonavir's alone. “Inhibitor” is not a universal rule; know the substrate and the booster.
- Induction lags. Starting or stopping an inducer like efavirenz takes one to two weeks to reach a new enzyme steady state — so the interaction, and its reversal, is delayed rather than immediate. A regimen switch is a psychotropic-level event even when nothing about the psychotropic changed.
- The traffic runs both ways. Antiretrovirals act on psychotropics, but psychotropic inducers act back on antiretrovirals — carbamazepine can lower dolutegravir enough to threaten virologic control. In HIV, your prescription can break their regimen.
The Interaction Logic
Renal dosing had a clean gate; HIV has a short sequence of questions that size the interaction. Three frame the pharmacokinetic side:
| Question | What you're asking | Why it matters |
|---|---|---|
| 1. Which bucket? | Is there a booster (/r, /c), an NNRTI inducer, or a clean integrase anchor? | Sets the direction before you look at any single drug. Booster → up; inducer → down; clean → probably neither. |
| 2. Is my psychotropic a CYP3A4 substrate? | Does the drug I want route through the enzyme the regimen touches? | A booster only matters if your drug is metabolized by the pathway it inhibits. Substrates swing; non-substrates largely escape. |
| 3. Does my psychotropic act back on the regimen? | Is the psychotropic itself an inducer (carbamazepine, oxcarbazepine, St John's wort, modafinil)? | This is the direction that causes virologic failure — the consequence is measured in resistance, not side effects. |
The fourth question sits outside the pharmacokinetic logic
- What do they add up to, regardless of level? That's not an interaction you dose around — it's an additive-risk decision. Additive QTc prolongation (antipsychotics plus certain antiretrovirals), additive sedation and CNS toxicity (efavirenz's neuropsychiatric load plus a sedating psychotropic), additive serotonergic tone, and a lowered seizure threshold in a population already at neurologic risk. The pharmacodynamic stack can harm even when every plasma level looks fine. Treat it as a separate decision on the same patient.
Which Psychotropics Need Caution (Orientation)
This chapter gives you the map, not the agent-by-agent interaction tables — those live in the member chapters. The HIV landscape sorts by interaction risk rather than by clearance, and the sort depends heavily on which bucket the regimen falls in.
| Tier | Agents (orientation only) |
|---|---|
| Highest concern contraindicated pairs or two-way traps |
Oral midazolam and triazolam with any booster — a labeled contraindication (excessive sedation, respiratory depression). Carbamazepine — the two-way collision: boosters push it toward toxicity, and it strips dolutegravir and other antiretrovirals enough to threaten control. Pimozide with boosters (QTc). St John's wort with essentially any regimen (induction → virologic failure). |
| Adjust / watch level shifts with the bucket |
Most antipsychotics (quetiapine, lurasidone, aripiprazole — levels rise on boosters; watch additive QTc), several SSRIs / SNRIs (paroxetine and others shift with boosted regimens), bupropion (the paradox — it falls on ritonavir-containing regimens), and the oxidatively-cleared benzodiazepines (alprazolam, diazepam) that climb with boosters. |
| Relatively clean few interactions either way |
On a modern integrase anchor, most agents behave close to normal. The conjugated benzodiazepines — lorazepam, oxazepam, temazepam (the “LOT” rule) — avoid the CYP3A4 tangle. Renally-cleared agents (lithium, gabapentin, pregabalin) largely sidestep the interaction axis — watch the kidney, and tenofovir's renal effects, instead. “Clean” is not “ignore”: the pharmacodynamic axis still applies. |
Two moves carry most of the benefit: identify the bucket before you choose the agent, and when you need a benzodiazepine reach for a conjugated one (LOT) that skips the CYP3A4 fight the boosters pick.
Modern vs Legacy Regimens: A Stable Target vs a Loaded One
The interaction burden of HIV therapy has fallen sharply, and the framework has to account for which era a patient's regimen belongs to. A modern first-line regimen — bictegravir/emtricitabine/tenofovir alafenamide, or a dolutegravir-based combination — is a relatively stable, low-interaction target: choose the psychotropic on its own merits, mind the one bidirectional trap (carbamazepine), and watch the pharmacodynamic stack.
A legacy or second-line regimen is the loaded one. A boosted protease inhibitor (darunavir/ritonavir), a cobicistat-boosted regimen, or efavirenz still turns up — in patients established years ago, in resistance, in specific clinical circumstances, and across much of the world. These are where the classic interactions live, and where the reflex to check every pairing is warranted. The practical instruction is to read the actual regimen every time rather than assume the modern default: the same diagnosis carries a very different interaction profile depending on the three letters after the slash. And when a regimen changes, re-run the interaction check — a switch off efavirenz can raise a psychotropic's level a week or two later without anyone touching the psychiatric prescription.
The Action Ladder
Integrase-anchored regimen (bictegravir / dolutegravir), CD4 preserved
The low-interaction default. Choose the psychotropic on its own merits and start at usual doses. Avoid carbamazepine (it strips the integrase inhibitor); watch the pharmacodynamic stack — QTc, sedation, serotonin.
Boosted regimen (/r or /c)
Assume CYP3A4 substrates will rise. Avoid oral midazolam / triazolam and pimozide; start sensitive agents low; remember bupropion falls with ritonavir (not with cobicistat). Favor conjugated benzodiazepines and renally-cleared agents that sidestep the enzyme.
Efavirenz / NNRTI regimen
Expect psychotropic levels to run low, and expect the antiretroviral's own neuropsychiatric burden. Distinguish drug-induced depression, insomnia, and vivid dreams from the primary disorder before escalating; re-check one to two weeks after any regimen change.
CD4 < 200 with new neuropsychiatric symptoms
Stop and rule out the CNS opportunistic infection first — image and, where indicated, tap before you attribute the picture to primary psychiatric illness. The pharmacology waits until the infection is excluded.
Clinical Pearls
Pearls
- Read the regimen, not the CD4 count, for interactions. Immune status tells you infection risk; only the named drugs tell you the drug interaction. Scan for a “/r” or “/c” and for efavirenz first.
- Two axes, not one. Pharmacokinetic level shifts and additive pharmacodynamic risk are separate decisions — a reassuring level doesn't clear an additive-QTc or additive-sedation problem.
- The booster paradox. Ritonavir inhibits CYP3A4 but induces UGT / CYP2B6, so bupropion drops on ritonavir-containing regimens instead of climbing — cobicistat lacks the induction, so bupropion is unchanged with a “/c” booster. Don't assume every substrate rises.
- Your prescription can break their regimen. Carbamazepine (and St John's wort, and other inducers) can lower antiretroviral levels enough to cause virologic failure and resistance — the interaction that costs more than a side effect.
- New psychosis at CD4 < 200 is an infection until proven otherwise. Toxoplasmosis and cryptococcus present psychiatrically; image and tap before you prescribe.
- A regimen switch is a psychotropic event. Starting or stopping efavirenz shifts levels over one to two weeks — re-check efficacy and tolerability after any antiretroviral change.
- Efavirenz has its own psychiatry. Vivid dreams, insomnia, depression, and a suicidality signal belong to the drug — consider the regimen before diagnosing a new primary disorder.
- The modern default is clean. On bictegravir or dolutegravir with no booster and no NNRTI, most psychotropics behave normally — treat the psychiatric illness properly rather than reflexively underdosing.
Red Flags — Stop and Reassess
Hold and look closer
- New psychosis, confusion, focal deficit, or personality change in someone with CD4 < 200 (or unknown) — image and consider LP for a CNS opportunistic infection before attributing it to primary psychiatric illness.
- Oral midazolam or triazolam ordered for a patient on a boosted (/r, /c) regimen — a labeled contraindication; stop and choose a conjugated benzodiazepine.
- Carbamazepine started, or already running, in a patient on dolutegravir or a boosted regimen — a two-way trap; verify with HIV pharmacy before continuing.
- Loss of antidepressant or antipsychotic efficacy after an antiretroviral switch — suspect induction (new or continuing NNRTI) stripping the level; re-check one to two weeks out.
- New depression, vivid dreams, or suicidal ideation coinciding with efavirenz — consider the antiretroviral as the driver and discuss a regimen change with the HIV team.
Patient Counseling Script
Plain-language script
“Your HIV medication and this psychiatric medication share the same processing system in the body, so they can change each other's levels — that's why I'm choosing this particular one and why we'll start carefully. Two things to know. First, never stop, start, or change your HIV medication without telling me, because a change on that side can quietly raise or lower this one over a week or two. Second, don't add any new medication, herbal product — St John's wort especially — or supplement without checking with us, because some of them can weaken your HIV treatment. And if you notice new low mood, strange dreams, or thoughts of harming yourself, call us — sometimes that comes from the medicines, and we can fix it. We work as one team with your HIV provider on all of this.”
EMR / Documentation Template
References
- Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the Use of Antiretroviral Agents in Adults and Adolescents with HIV. U.S. Department of Health and Human Services; September 2025 update. (Integrase-inhibitor-based regimens as first-line; antiretroviral drug-interaction tables.)
- National HIV Curriculum (University of Washington). Drug Interactions with Antiretroviral Medications. hiv.uw.edu; 2026. (Bictegravir and dolutegravir have relatively few interactions; NNRTI induction and booster inhibition profiles.)
- Ritonavir (Norvir) Prescribing Information. AbbVie. (Potent CYP3A4 inhibition; contraindication with oral midazolam and triazolam; UGT / CYP2B6 induction underlying the bupropion decrease.)
- Song I, Weller S, Patel J, et al. Effect of carbamazepine on dolutegravir pharmacokinetics and dosing recommendation. J Clin Pharmacol. (Enzyme induction lowers dolutegravir exposure; dose adjustment required.)
- Clinical implications and management of drug-drug interactions between antiretroviral agents and psychotropic medications. Ment Health Clin. 2013;2(9):274. (Paroxetine and bupropion AUC reductions with boosted regimens; class-by-class psychotropic–ART interactions.)
- Efavirenz (Sustiva) Prescribing Information. Bristol-Myers Squibb. (Neuropsychiatric adverse effects — depression, suicidality, sleep disturbance; CYP induction.)
- HIV-1–Associated Opportunistic Infections. StatPearls. NCBI Bookshelf; 2026. (CD4-stratified CNS opportunistic infections — toxoplasmosis, cryptococcal meningitis, PML, CMV — and their neuropsychiatric presentations.)
Last reviewed July 2026. Sensitive-content note: references suicidality as an antiretroviral adverse effect (adverse-effect context only). Part of the Psychiatry Education Forum Academy; for clinician education — it supports, and does not replace, individual clinical judgment and current local protocols.
You have the framework. Now get the drug-by-drug answers.
This chapter taught you how to decide which way an antiretroviral bends your psychotropic — and how to keep the level question separate from the additive-risk question. The member chapters give you the agent-by-agent calls, starting with the class you reach for most and the interaction that surprises everyone.
Educational use only. Refer to the sources cited above and current prescribing information for clinical decisions. Psychiatry Education Forum and authors assume no liability for use of this material.
Responses