D1/D2/D3/D4 dopamine agonists? (eg Dihydrexidine)

#1

What are ways to get them (if possible).

They may have some of the benefits of stimulants, more precision, and way less of the neurotoxicity esp if you’re NQO1 P187S

Dopamine receptor D1 - Wikipedia [this has existed for a long time]…

1 Like
#2

They have a higher risk of hallucinations and psychosis than other drugs that act around dopamine.

Originally synthesized from ergot, the same as the semi-synthetic LSD.

#3

Claude Sonnet 5 summary of the different dopamine receptors and their general actions when agonised (keeping in mind that different agonists can be exceptions to this rule somewhat):

D1-like family: D1 (DRD1) and D5 (DRD5)

  • Couple to Gs (sometimes Golf in striatum for D1)
  • Agonism → activates adenylyl cyclase → increases cAMP → activates PKA
  • Generally excitatory/depolarizing effect on postsynaptic neurons
  • D1 is the most abundant dopamine receptor in the brain, dense in striatum, nucleus accumbens, prefrontal cortex
  • D5 is far less abundant, expressed in hippocampus and hypothalamus, has roughly 10x higher affinity for dopamine than D1 but no clinically distinct agonists exist to exploit that

D2-like family: D2 (DRD2), D3 (DRD3), D4 (DRD4)

  • Couple to Gi/Go
  • Agonism → inhibits adenylyl cyclase → decreases cAMP; also opens GIRK potassium channels and inhibits voltage-gated Ca2+ channels
  • Generally inhibitory/hyperpolarizing effect
  • D2 exists as two splice variants: D2short (mostly presynaptic autoreceptor, inhibits dopamine synthesis and release via negative feedback) and D2long (mostly postsynaptic). D2 is the primary target of essentially all antipsychotics (as antagonists) and of dopamine agonists used in Parkinson’s disease
  • D3 is concentrated in limbic areas (nucleus accumbens, islands of Calleja) rather than motor striatum, implicated in reward/motivation circuitry, and has become a target of interest for addiction pharmacology because it’s less involved in motor side effects
  • D4 has the most genetic variability of the five (notably a variable-number tandem repeat in exon 3), is expressed at low density mostly in frontal cortex, amygdala, hippocampus, and has historically been linked (weakly, with mixed replication) to novelty-seeking personality traits and ADHD

Net functional picture when agonised:

  • D1-like agonism raises intracellular cAMP → excitatory postsynaptic signaling, involved in working memory (D1 in prefrontal cortex) and direct-pathway striatal output (“go” signal in basal ganglia)
  • D2-like agonism lowers cAMP → inhibitory signaling, involved in indirect-pathway striatal output (“no-go” signal), autoreceptor-mediated feedback inhibition, and is the mechanism exploited by drugs like pramipexole/ropinirole in Parkinson’s

The 5 dopamine receptor types and their agonists:

Claude Sonnet 5 summary of all dopamine receptor agonists listed in these Wikipedia articles:

Dopamine Receptor Agonists (D1–D5): Effects, Status, and Availability

Compiled from the Wikipedia agonist listings for DRD1–DRD5, cross-referenced. Many D2-like agonists hit D2, D3, and D4 simultaneously (they’re not receptor-clean tools), so entries are cross-listed where relevant rather than force-fit into one bucket.

D1-like family (D1 + D5): Gs-coupled, ↑cAMP, generally excitatory

D1 receptor agonists

Clinically used, but non-selective (D2-preferring, incidentally hit D1):

  • Apomorphine — broad-spectrum dopamine agonist (D1, D2, D3, D4 all hit), used clinically for Parkinson’s “off” episodes (subcutaneous injection/sublingual film). Strong emetic effect (used historically as an emetic), can cause hypotension, sedation, compulsive behaviors (impulse control disorders) with chronic use. Available by prescription.
  • Pergolide — ergot-derived, D1/D2 agonist, was used for Parkinson’s/RLS but withdrawn from most markets (2007, incl. US) due to cardiac valvulopathy risk (5-HT2B agonism). Still available in some countries for animal use (veterinary, Cushing’s in horses).
  • Rotigotine — transdermal patch, D1/D2/D3 agonist, approved for Parkinson’s and restless legs syndrome. Non-ergot, better cardiac safety profile than ergot agonists. Widely available by prescription.
  • Terguride — partial agonist, ergot derivative, used in some countries for hyperprolactinemia; not licensed in the US.

Selective D1-like agonists (research/clinical trial only, none approved for CNS indications):

  • Fenoldopam — selective D1 partial agonist, does NOT cross the blood-brain barrier. Approved and available as an IV antihypertensive (used in hospital settings for hypertensive emergencies). This is the one D1 agonist with genuine mainstream clinical use, precisely because it’s peripherally restricted.
  • Dihydrexidine — first full, selective D1 agonist demonstrated in humans. Phase IIa trials as a cognitive enhancer (schizophrenia, Parkinson’s). Showed real antiparkinsonian efficacy in primate models and increased prefrontal perfusion in schizophrenia patients at 20mg. Dose-limiting side effect: profound hypotension. Short half-life and poor oral bioavailability killed further development. Never reached market.
  • A-86929 / Adrogolide (ABT-431, prodrug) — full D1 agonist, 14-fold D1 selectivity. Adrogolide (the prodrug, improved bioavailability) was studied clinically for Parkinson’s. Also stalled at hypotension/pharmacokinetic hurdles.
  • Dinapsoline, Dinoxyline, Doxanthrine — full agonists, research tools only (animal data), never in human trials.
  • SKF-81297, SKF-82958, SKF-38393 — classic pharmacology research tools (benzazepine series), animal data only, not developed as drugs.
  • 6-Br-APB, Trepipam (SCH-12679) — research compounds.
  • Tavapadon (CVL-751) — non-catechol, orally bioavailable, biased D1/D5 partial agonist. This one is notable: it’s in active Phase 3 clinical development for Parkinson’s disease (Cerevel/AbbVie), the most clinically advanced selective D1-class agonist in decades, specifically engineered to sidestep the old hypotension/bioavailability problems. Not yet approved as of the most recent data I have — worth checking current trial status if this matters for your piece.
  • Razpipadon (CVL-871) — related Cerevel D1/D5 agonist program, being explored for apathy in Alzheimer’s/dementia, not psychostimulant-relevant.
  • Stepholidine — a natural alkaloid (from Corydalis/Stephania plants) with the unusual profile of D1 agonist + D2 antagonist. Used in Chinese traditional medicine contexts; some human data suggesting antipsychotic-like effects. Not FDA-approved.
  • Clozapine — mainly known as a D2/5-HT2A antagonist antipsychotic, but has partial D1 agonist activity as a secondary property. Widely available (prescription, heavily monitored due to agranulocytosis risk).

Bottom line for D1: nothing centrally-acting and D1-selective has ever reached market. The clinical experience (dihydrexidine, adrogolide) says D1 agonism gives real cognitive/motor benefits but hypotension is a hard ceiling — this is relevant if you’re evaluating “safer than stimulants,” since orthostatic hypotension is its own real-world hazard (fainting, falls).

D5 receptor agonists

There are no known D5-selective agonists. D5 is ~49-80% homologous to D1 depending on the domain, and every known agonist that hits D1 also hits D5 (they can’t be pharmacologically separated with current ligands). So functionally, “D5 agonism” in practice just means “whatever D1 agonist you’re using is also going to hit D5.” D5 has ~10x higher affinity for dopamine itself than D1, and shows high constitutive (agonist-independent) activity — it’s “on” a bit even without a ligand. A selective D5 antagonist exists (4-chloro-7-methyl-…hexahydrodibenz[d,g]azecin-3-ol) but no selective agonist. Not clinically relevant for a stimulant-alternative screen since you can’t dose D5 without D1.

D2-like family (D2, D3, D4): Gi/Go-coupled, ↓cAMP, generally inhibitory

D2 receptor agonists

Approved, currently marketed (mostly for Parkinson’s/hyperprolactinemia/RLS):

  • Bromocriptine — ergot-derived, D2 full agonist. Approved for Parkinson’s, hyperprolactinemia, and (as Cycloset) type 2 diabetes glycemic control. Ergot-related valvulopathy/fibrosis risk at Parkinson’s-range doses; the diabetes dose is much lower and considered safer on that front. Widely available generically.
  • Cabergoline — ergot-derived, long half-life, mainly used for hyperprolactinemia (prolactinomas) and off-label RLS/Parkinson’s. Same ergot valvulopathy concern, dose-dependent. Still on-market, prescription.
  • Pramipexole — non-ergot, hits D2 AND D3 (arguably D3-preferring — see below), and D4. Approved for Parkinson’s and RLS. No valvulopathy risk (non-ergot), but strong association with impulse control disorders (compulsive gambling, shopping, hypersexuality, binge eating) — this is a well-documented, dose-related class effect tied to D3 activation in limbic reward circuitry. Widely prescribed.
  • Ropinirole — non-ergot, D2 full agonist (also D3-active). Approved Parkinson’s/RLS. Same impulse-control-disorder signal as pramipexole. Widely prescribed.
  • Rotigotine — see D1 section above; also full D2/D3 agonist, transdermal, same ICD risk class-wide.
  • Quinagolide — non-ergot D2 agonist for hyperprolactinemia, available in Europe/elsewhere, not marketed in US.
  • Talipexole — used in Japan for Parkinson’s, not in US/EU markets.
  • Aripiprazole, Brexpiprazole, Cariprazine — atypical antipsychotics that act as D2 partial agonists (this is their defining mechanistic feature vs. pure antagonist antipsychotics). Widely prescribed for schizophrenia, bipolar disorder, depression augmentation. Not stimulant-like in effect — net effect is usually calming/antipsychotic, though they can occasionally provoke impulse-control issues (aripiprazole specifically has an FDA warning for compulsive behaviors, same D2/D3 mechanism as the Parkinson’s drugs).

Interesting incidental D2 partial agonists (approved for other indications):

  • Modafinil / Armodafinil — marketed as wakefulness agents (narcolepsy, shift-work disorder, OSA-related sleepiness). Long thought to work purely via weak dopamine transporter (DAT) inhibition, but there’s evidence armodafinil is also a D2 partial agonist at “D2High” (activated-state) receptors. This is one of the more directly relevant entries for your project: modafinil is already widely regarded as a “gentler” stimulant-alternative with lower abuse liability than amphetamines, and this D2 mechanism may be part of why. Prescription-only, Schedule IV in the US.
  • Buspirone — anxiolytic; antagonist at higher doses but there’s literature describing autoreceptor-level D2 modulation at lower doses. Widely available, low abuse potential.

Not approved / experimental / research tools:

  • N,N-Propyldihydrexidine, Quinelorane, Quinpirole, Sumanirole, Aplindore — research compounds, animal pharmacology only (sumanirole is notable as a highly D2-selective full agonist used to dissect D2 vs D3 contributions in lab studies, never developed clinically).
  • Piribedil — approved and marketed in parts of Europe (France, etc.) for Parkinson’s and some vascular/cognitive indications, but not FDA-approved in the US.

Recreational/illicit substances with reported D2 activity (context, not a recommendation):

  • LSD — in vitro partial D2 agonist, potentiates prolactin secretion; primary mechanism is 5-HT2A agonism. Schedule I in most jurisdictions.
  • Salvinorin A — primarily a κ-opioid agonist; D2 partial agonism is a minor, secondary property. Legal status varies by jurisdiction (unscheduled federally in the US, but banned in some states).
  • Ketamine, Memantine — primarily NMDA antagonists; D2High agonist activity described as a secondary mechanism (Seeman’s work). Ketamine is a controlled substance/anesthetic; memantine is an approved Alzheimer’s drug, freely prescribable, low abuse liability.
  • Cannabidiol (CBD) — listed as having some D2 interaction; widely available, not scheduled federally in the US (hemp-derived), low abuse potential, but the D2 mechanistic contribution here is not well characterized and CBD’s psychoactive/stimulant-like effects are minimal to none.

D3 receptor agonists

D3 has the standout feature of highest binding affinity for dopamine of any subtype, concentrated in limbic areas (nucleus accumbens, islands of Calleja) rather than motor striatum — this is why it’s the receptor most tied to reward/motivation/addiction circuitry rather than movement.

  • Pramipexole, Ropinirole, Rotigotine — as above; these are actually more potent/selective at D3 than at D2 in receptor-binding terms, even though they’re marketed as “D2 agonists.” The impulse-control-disorder signal associated with this drug class is generally attributed mechanistically to D3 activation in the mesolimbic reward pathway, not D2 in the nigrostriatal motor pathway. This is a genuinely important point for a “safer stimulant” screen: D3 agonism looks like exactly the mechanism you’d want to avoid if the goal is avoiding compulsive/addictive reinforcement patterns.
  • 7-OH-DPAT — a D3-preferring agonist, research tool only (no clinical development), used in animal studies including antidepressant-like effect models in rodents (olfactory bulbectomy paradigm).
  • Apomorphine — as above, non-selective, hits D3 among others.
  • Partial agonists at D3: Aripiprazole, Brexpiprazole, Cariprazine (same antipsychotics as above — cariprazine in particular is notably D3-preferring over D2, which is thought to underlie some of its distinct pro-cognitive/anti-anhedonic clinical signal vs other antipsychotics).

D3 antagonists (not what you asked, but relevant context) are an active area of addiction-pharmacology research — e.g., SB-277011-A reduces drug-seeking behavior in animal models — because blocking D3 seems to blunt the reinforcing/craving side of reward circuits. That’s the mirror image of what agonism does, and underscores that D3 agonism is probably the least “safe target” of the five if compulsive-use liability is your concern.

D4 receptor agonists

D4 is the outlier of the D2-like family: sparse expression (mostly frontal cortex, amygdala, hippocampus — not the striatum where movement and classic reward signaling live), highly polymorphic (the exon 3 VNTR, with the 7-repeat allele weakly/inconsistently linked to ADHD and novelty-seeking), and its agonists have a distinctive effect profile centered on working memory and cognitive enhancement rather than movement or classic reward.

None of these are approved drugs. All human/clinical data is thin to nonexistent; most of what exists is rodent/primate.

  • A-412,997 — full agonist, highly selective (>100-fold over 70 other receptors/channels tested). In rats: improved cognitive performance, increased locomotor activity, notably without affecting reward-related behavior — i.e., it looks pro-cognitive without the addictive-reinforcement signature you’d worry about with a stimulant. This is arguably the most interesting single data point in the whole D4 literature for your specific research angle, but it’s never been in a human trial as far as the record shows.
  • ABT-724 / ABT-670 — developed by Abbott specifically for erectile dysfunction (a legacy of dopamine’s role in sexual function via D4). ABT-670 has better oral bioavailability. Neither reached market; ED drug development in this class was abandoned in favor of PDE5 inhibitors (sildenafil etc.).
  • WAY-100635 — technically a potent D4 full agonist, but this compound is much better known and used in neuroscience as a 5-HT1A antagonist; the D4 activity is a secondary/incidental finding, not why anyone uses it.
  • FAUC 316, FAUC 299, PIP3EA, Ro10-5824, F-15063 — partial agonists, research tools; F-15063 was explored as an antipsychotic candidate with partial D4 (and D2/D3) agonism, doesn’t appear to have reached market.
  • Flibanserin — this one is actually approved and marketed (Addyi, for hypoactive sexual desire disorder in women), but its D4 partial agonism is a minor part of a complex multi-receptor profile (mainly 5-HT1A agonist/5-HT2A antagonist). Carries an alcohol-interaction black-box warning (hypotension/syncope) and modest efficacy — not something anyone would repurpose for cognitive/stimulant purposes.
  • Roxindole, Apomorphine, Nuciferine — non-selective, hit D4 alongside D2/D3 (nuciferine, a lotus-plant alkaloid, is D4-selective but also partial D2/D5 agonist — mostly known from recreational “blue lotus” folklore, no rigorous human pharmacokinetic/safety data exists).

Notable dose-response quirk: D4 agonists in animal studies show a U-shaped dose-response curve for cognitive/working-memory benefits — too little or too much and the benefit disappears or reverses. That non-monotonic relationship is a real practical complication for any drug-development angle (narrow therapeutic window), separate from any question of safety per se.

Cross-cutting notes relevant to a “stimulant alternative” screen

  1. The genuinely safest-feeling mechanism in this whole list, based on existing human data, is peripheral D1 agonism (fenoldopam) — but it doesn’t cross the blood-brain barrier, so it has zero cognitive/energizing effect. It’s a cautionary example: receptor-selectivity doesn’t help you if the molecule can’t get where you want it.
  2. Central D1 agonism (dihydrexidine/adrogolide/tavapadon) has the best-documented pro-cognitive signal of the five subtypes in actual human trials, but the historical dose-limiting toxicity was hypotension, not addiction liability — worth noting since “safer than stimulants” could mean either “non-habit-forming” or “no cardiovascular/psychiatric side effects,” and D1 agonism’s problem is specifically the latter.
  3. D2 agonism is the most clinically mature of the five (decades of Parkinson’s/RLS use), but the entire non-ergot class (pramipexole, ropinirole, rotigotine) carries a well-established, dose-dependent impulse-control-disorder signal. That’s arguably a bigger real-world “addiction-adjacent” liability than anything associated with modafinil.
  4. D3 agonism is mechanistically the one most tied to reward/reinforcement circuitry specifically (not movement), which makes it the least promising direction if your goal is a lower-abuse-liability profile — the pharmacology point wouldn’t support pitching a D3 agonist as “safer,” even though several D3-active drugs are already marketed for other purposes.
  5. D4 agonism is the most under-explored and, on the rodent data, the most interesting for a “cognitive enhancement without reward-circuit activation” narrative — A-412,997 in particular. But there is essentially no human safety or efficacy data, so any framing has to stay clearly in the “promising unexplored target” register, not “here’s a safer stimulant that exists.”
  6. Modafinil/armodafinil is really the closest thing that already exists on the market to what you may be describing — a stimulant-class-adjacent wakefulness agent with a comparatively low abuse-liability profile, and part of its mechanism (D2 partial agonism at the “high-affinity state” receptor pool) is directly from this literature. Worth treating as the real-world benchmark/case study rather than a hypothetical.
#4

5.6-sol-medium:

1. Tavapadon: strongest D1/D5 asset, but its trial outcomes are already known

Tavapadon—formerly PF-06649751 or CVL-751—is AbbVie’s oral, once-daily, selective dopamine D1/D5 partial agonist for Parkinson’s disease. Endpoint Arena already includes it in the AbbVie pipeline as:

Tavapadon — D1/D5 partial agonist — Parkinson’s Disease — Submitted

The pivotal trials include:

Trial NCT Population Status for prediction
TEMPO-1 NCT04201093 Early Parkinson’s, fixed-dose monotherapy Results public; primary endpoint positive
TEMPO-2 NCT04223193 Early Parkinson’s, flexible-dose monotherapy Results public
TEMPO-3 NCT04542499 Adjunctive to levodopa in fluctuating Parkinson’s Results public; primary endpoint positive
TEMPO-4 NCT04760769 Long-term open-label extension Primarily long-term safety; poor binary efficacy-market candidate

TEMPO-1 showed statistically significant improvement in the combined MDS-UPDRS Parts II and III score at Week 26, and TEMPO-3 showed an approximately 1.1-hour placebo-adjusted increase in daily “on” time without troublesome dyskinesia. Those results are already public in the TEMPO-1 publication and TEMPO-3 results.

AbbVie submitted the tavapadon NDA in September 2025. The company describes it as seeking use both as monotherapy and as adjunctive therapy to levodopa. AbbVie’s official NDA announcement confirms that submission.

Endpoint Arena verdict for tavapadon

  • Excellent drug/pipeline candidate.
  • Poor candidate for a new clinical-trial endpoint market because the pivotal outcomes are already known.
  • Potentially excellent candidate for an FDA approval market, if Endpoint Arena accepts regulatory-decision questions.
  • The public nomination form nevertheless requires an NCT number, so the closest supporting identifier would be NCT04201093, while clearly stating that the unresolved event is FDA approval—not whether TEMPO-1 succeeded.
  • Endpoint Arena may reject that nomination if its market policy is strictly limited to prospective trial readouts.

A possible submission would be:

Consider a regulatory decision market for AbbVie’s tavapadon, a selective dopamine D1/D5 partial agonist submitted to the FDA for Parkinson’s disease as monotherapy and adjunctive therapy. Supporting pivotal trial: NCT04201093, with additional evidence from NCT04223193 and NCT04542499. The Phase 3 efficacy results are already public, so the unresolved event should be FDA approval of tavapadon—not whether the completed trial met its endpoint. Tavapadon could become the first approved selective D1/D5 partial agonist for Parkinson’s disease.

I would not describe any exact FDA action date unless AbbVie or FDA has officially disclosed it. Public secondary sources conflict about the expected date.

2. Glovadalen: probably the best future Endpoint Arena trial candidate

Glovadalen, also called UCB0022, is UCB’s oral dopamine D1 receptor positive allosteric modulator. Unlike tavapadon, it does not directly activate D1 receptors indiscriminately; it is intended to amplify the response to dopamine where and when dopamine is present.

Its completed Phase 2a ATLANTIS study was:

  • Trial: NCT06055985
  • Sponsor: UCB
  • Population: advanced Parkinson’s disease with motor fluctuations
  • Primary endpoint: change from baseline through Day 70 in average daily “off” time
  • Status: completed
  • Outcome: UCB has reported the study as positive

The completed study is documented on UCB’s glovadalen clinical-study page. UCB now says a larger signal-confirming Phase 2 study is planned to begin in the first half of 2027, according to its current investor pipeline.

That future study is the strongest D1-related candidate for Endpoint Arena because:

  • The result will be genuinely prospective.
  • It should have a randomized control and prespecified efficacy endpoint.
  • Parkinson’s “off” time is understandable, patient-relevant, and objectively resolvable from the statistical analysis.
  • D1 positive allosteric modulation is mechanistically differentiated from established D2/D3 agonists.
  • The prior positive Phase 2a provides enough evidence for a meaningful forecast without making the result foregone.

The limitation is that the signal-confirming study does not yet appear to have a public NCT identifier. Endpoint Arena’s suggestion form requires an exact NCT number, so it cannot be properly nominated until the study is registered.

Once registered, the submission should say something like:

Please add UCB’s signal-confirming Phase 2 study of glovadalen, an oral dopamine D1 receptor positive allosteric modulator, in Parkinson’s disease. The earlier ATLANTIS study, NCT06055985, reportedly met its primary endpoint by reducing average daily OFF time. The new randomized study will prospectively test whether that finding replicates in a larger population. The market should resolve YES only if the prespecified glovadalen dose or pooled treatment comparison achieves statistical significance versus placebo on the new trial’s protocol-defined primary endpoint.

The new NCT number—not NCT06055985—should go in the nomination form.

Other D1 candidates

Mevidalen (LY3154207) is a selective D1 positive allosteric modulator with earlier Parkinson’s and Lewy body dementia studies. However, I did not find a sufficiently current, unresolved later-stage trial that would compete with glovadalen as an Endpoint Arena candidate.

PF-06412562 produced early feasibility data in advanced Parkinson’s disease, but it does not currently present a clear unresolved, registered pivotal catalyst.

Recommendation

The clean ranking is:

  1. Glovadalen’s upcoming signal-confirming Phase 2 trial — best prospective clinical-endpoint market; wait for the new NCT registration.
  2. Tavapadon FDA approval — strongest near-term D1 commercial prediction, but Endpoint Arena may not support regulatory-only markets.
  3. Tavapadon TEMPO trials — do not nominate as new predictions; their important efficacy results are already public.
  4. Mevidalen or PF-06412562 — no compelling currently unresolved trial catalyst located.

So, if the goal is to get a dopamine D1 prediction onto Endpoint Arena, I would watch for UCB’s new glovadalen trial registration and nominate it immediately. Tavapadon is scientifically and commercially more advanced, but its remaining uncertainty is regulatory rather than whether its Phase 3 efficacy trials succeeded.

#5

Why the interest in pan D receptor agonists? If you’re just trying to replace stimulants for ADHD treatment, D2 selective agonists are going to be the closest target for the ADHD receptor deficit. Of the D2 agonists, cabergoline has best risk/benefit profile.

#6

https://www.nature.com/articles/s41386-026-02540-w

Nqo1…

#7

A functional D1 agonist would definitely be useful, but all the historically developed ones have major issues and as such I would avoid. Tavapadon and Glovadalen are interesting. Until they’re approved somewhere, sourcing would be difficult. Tavapadon is active in the prefrontal cortex, so would be a good candidate for ADHD. However the receptor desensitization studies were done with striatal neurons, so whether the same behavior occurs in prefrontal cortex neurons is unproven, but likely given the receptor binding mechanism. Glovadalen’s potential for ADHD is harder to estimate given the limited data published on it and what has been published is focused on the striatus.

1 Like
#8

D2 is what one prof I emailed suggested

#9

MAO-A inhibitors would likely be very effective, however anyone that is considering trying one needs to be aware of the potential for serotonin syndrome and the tyramine pressor response which can be fatal in extreme cases.