Prescribing Duloxetine

Why Duloxetine Earns Its Place

Duloxetine is the SNRI you reach for when a patient's depression or anxiety travels with a body complaint (neuropathic pain, fibromyalgia, aching joints, chronic low back pain) or when you want the noradrenergic reach of an SNRI without venlafaxine's blood-pressure liability. It is the most broadly FDA-labeled antidepressant in psychiatry, carrying six approved indications spanning mood, anxiety, and three distinct pain syndromes. It is generic, once-daily, easy to titrate, and mechanistically boring in the good way: it blocks serotonin and norepinephrine reuptake and does nothing else at the receptor level.

But sell it honestly. The "Depression Hurts" marketing built a reputation for painful depression that the data never earned: the analgesic effect size in depression is roughly 0.115 (p≈0.057, not significant), and much of what patients feel is placebo. Where duloxetine genuinely delivers is narrower and more useful than the ad campaign claimed: GAD (where a Lancet network meta-analysis ranked it among the most efficacious and best-tolerated agents on the highest-quality evidence), diabetic neuropathic pain and fibromyalgia (real, FDA-backed, though tricyclics often match or beat it), and depression in patients you want to keep off a drug that raises blood pressure.

A note on the name: despite "SNRI," duloxetine is roughly 20:1 selective for serotonin over norepinephrine. It is more balanced than venlafaxine (120:1), but it is still, functionally, a serotonergic drug with a modest noradrenergic tail. Do not expect the "norepinephrine magic" (energy, motivation, focus) that the class name implies. The evidence for selective noradrenergic benefit does not exist.

Part 1: Indications: Who Is Duloxetine For?

FDA-Approved Uses

• Major depressive disorder (MDD)
• Generalized anxiety disorder (GAD), including a pediatric GAD indication (ages 7+)
• Fibromyalgia
• Diabetic peripheral neuropathic pain
• Chronic low back pain
• Chronic musculoskeletal / osteoarthritis pain

Off-label: stress urinary incontinence (an approved indication in Europe and elsewhere, not in the US) and, weakly, OCD (sertraline is the better serotonergic choice there).

Where Duloxetine Actually Shines

Depression with comorbid anxiety: a top-tier choice

Duloxetine holds an FDA GAD indication, and in the 2019 Lancet network meta-analysis of GAD pharmacotherapy it ranked among the most efficacious agents and among the best tolerated, on the highest-quality evidence of any antidepressant examined. When a patient's depression and anxiety are tangled together, this is a rational, evidence-backed first move. (Well-tolerated GAD comparators from the same analysis: escitalopram, venlafaxine XR, pregabalin. Quetiapine XR was most effective but poorly tolerated.)

Depression with a comorbid pain syndrome

This is duloxetine's marketing home turf, and the honest version is more modest than the campaign:

• Diabetic neuropathic pain: the NNT is about 5.2. Real, FDA-approved, but weaker than tricyclics or venlafaxine (NNT ≈ 3). For practical purposes duloxetine and venlafaxine are interchangeable here.
• Fibromyalgia: only the 120 mg dose separates cleanly from placebo (40.1% response at 120 mg vs 34% at 60 mg vs 23.7% placebo in a 520-patient, 6-month RCT). Rheumatologists still reach for amitriptyline first, given its longer track record and lower cost.
• Pain within depression (the "Depression Hurts" claim): the benefit is marginal, effect size 0.115, under a 1-point advantage on the Brief Pain Inventory, and no superiority over paroxetine across five head-to-head trials. Placebo delivers much of the apparent analgesia.

The hypertensive or cardiovascularly cautious patient

Duloxetine's standout structural advantage over venlafaxine: it does not meaningfully raise blood pressure. Venlafaxine causes dose-dependent diastolic hypertension (watch especially ≥225 mg/day); duloxetine carries no such warning. If you want SNRI pharmacology in someone with hypertension or cardiovascular risk, duloxetine is the SNRI to pick.

Melancholic depression

SNRIs are the reasonable second-line here when a tricyclic can't be tolerated: the early-morning awakening, anorexia, psychomotor, ruminative-guilt phenotype that historically favored dual-mechanism drugs. This is clinical inference more than hard trial data, but duloxetine and venlafaxine are the sensible SNRI choices.

Treatment-resistant depression (as a switch)

A standard second- or third-step move: after an SSRI fails, switching to an SNRI is a mainstream option. Be honest about the ceiling: STAR*D taught us that switching classes buys little; tolerability and side-effect fit, not mechanism, drive the real-world decision.

Where Duloxetine Is Not the Answer

• Uncomplicated depression: an SSRI or bupropion is cheaper, at least as effective, and (for bupropion) sexually neutral and weight-favorable. Duloxetine has no efficacy edge over SSRIs for depression alone.
• Liver disease or heavy alcohol use: the hepatotoxicity signal makes this the wrong SNRI (see Part 2 and Part 6).
• Urinary retention risk (e.g., BPH): the noradrenergic effect on the bladder sphincter can worsen retention, paradoxical given its overseas incontinence indication, but a real caution.

Part 2: Before You Start: Workup and Candidacy

Duloxetine needs far less workup than lithium, but it is not a "no-labs" drug the way an SSRI is. The one screen that matters is the liver.

Baseline assessment

No routine ECG, CBC, or electrolytes are required. Renal function is worth knowing (dose-adjust in severe impairment), but duloxetine has no meaningful renal metabolism.

Who is a poor candidate?

• Liver disease or heavy alcohol use: avoid; use desvenlafaxine instead (glucuronidated, liver-sparing).
• Substantial urinary retention risk (BPH, prior retention).
• Narrow-angle glaucoma: rare but real; the noradrenergic effect can raise intraocular pressure.
• On an MAOI or within the washout window.
• Medically complex / heavy polypharmacy: the CYP2D6 inhibition and hepatic metabolism make desvenlafaxine the cleaner SNRI.

Part 3: How to Start and Dose

Formulations

Duloxetine comes as delayed-release capsules: 20, 30, 40, and 60 mg. It is generic and inexpensive. Take it with food or a snack, which aids absorption and blunts nausea. Once daily is standard; splitting BID (e.g., 30 mg twice daily) is a legitimate tolerability tool for nausea or dizziness.

Starting dose and titration, by indication

Depression and GAD

• Start 30 mg daily × 1 week, then 60 mg daily. (Starting at 30 mg first is the nausea-sparing move; you can start at 60 mg in the robust patient.)
• Target: 60 mg/day. This is the therapeutic dose.
• Max: 120 mg/day, but know that 120 mg is no more effective than 60 mg for depression or GAD. Escalating to 120 mg for mood or anxiety chases side effects, not benefit. Don't do it reflexively.

Fibromyalgia

• 60–120 mg/day. Here 120 mg does earn its keep: it's the only dose that separated cleanly from placebo (40.1% vs 23.7%). Start low (30 mg), titrate to 60, then to 120 if response is partial.

Diabetic neuropathy / chronic pain

• Standard 60 mg/day; some go to 120 mg. NNT ≈ 5.2 for diabetic neuropathy.

Titration compared to venlafaxine

Duloxetine's schedule is simpler than Effexor XR's, with fewer dose steps and an easier climb to a therapeutic dose. This is a genuine, if modest, quality-of-life advantage for both prescriber and patient.

Special titration situations

• Elderly: start low (30 mg or lower), go slow.
• Poor CYP2D6 metabolizers (genetic, or on a strong 2D6 inhibitor): expect higher levels; titrate more cautiously.
• Smokers/cannabis users: CYP1A2 induction lowers duloxetine levels, relevant if they quit (levels rise; see Part 7).

Part 4: Monitoring

Duloxetine is a light-monitoring drug compared with lithium, but it is not a zero-monitoring drug.

No routine ECG, CBC, electrolytes, or drug levels are needed.

Thresholds to act on

• ALT rising / clinical hepatitis signs: stop duloxetine and evaluate. Do not wait for jaundice.
• New BP elevation: less expected than with venlafaxine, but check it if the patient reports symptoms.
• Hyponatremia symptoms (confusion, lethargy, falls, especially in the elderly): SNRIs, like SSRIs, can cause SIADH; check sodium if suspected.

Part 5: Side Effects and How to Manage Them

The governing principle mirrors the SSRIs: most duloxetine side effects are early, dose-related, and either self-limited or manageable, but the ones that make patients quit are nausea early on, sexual dysfunction long-term, and the discontinuation syndrome at the end. Start low, titrate into the therapeutic dose, and manage proactively.

Gastrointestinal: nausea (the big early one)

Nausea is the most common early complaint and the most common reason for early dropout. It is dose-related and usually resolves within 2–4 weeks.

Management:

• Take with food.
• Start at 30 mg, not 60, and titrate.
• Split the dose (30 mg BID) if needed.
• Reassure: this almost always fades. Pushing through the first two weeks is usually the whole battle.

Sexual dysfunction

Treatment-emergent sexual dysfunction (TESD) is common, and, as with all SSRIs/SNRIs, the directly-asked rate is far higher than what patients volunteer. In head-to-head data, duloxetine ran about 46% TESD vs paroxetine 61% (paroxetine worst) vs placebo ~29% over 8 weeks in previously unaffected patients; over a 34-week extension the rates converged (~39% duloxetine vs ~45% paroxetine vs ~35% placebo). Bottom line: duloxetine sits mid-pack, clearly better than paroxetine, roughly comparable to escitalopram, and no better than other SNRIs.

Management (in order of evidence):

• Switch to a sexually neutral agent if this is a dealbreaker: bupropion (best, no sexual side effects), mirtazapine, or trazodone. Do not switch to another SNRI, as there's no advantage.
• PDE-5 inhibitors (sildenafil 50–100 mg) are robustly effective for antidepressant-induced ED in men and, in some data, women.
• Bupropion augmentation (SR 150–300 mg/day): inconsistent data, marginal, but a reasonable add-on if you want to keep the duloxetine.
• Vortioxetine has a slight TESD edge at low doses but not at therapeutic ones, not worth the cost/nausea trade for this alone.

Excessive sweating (diaphoresis)

A noradrenergic side effect, less prominent than with venlafaxine or levomilnacipran, but real.

Management:

• Terazosin 1–2 mg/day (alpha-blocker) or oxybutynin 5 mg 1–2×/day (anticholinergic).
• Glycopyrrolate off-label, but warn: anticholinergic sweat suppression impairs thermoregulation, a heat-stroke risk in hot conditions.

Dry mouth

Modest anticholinergic-like effect.

Management: OTC saliva substitutes, Biotene toothpaste/rinse, sugar-free gum. Counsel on increased caries risk; recommend regular dental care.

Constipation

From the noradrenergic tone.

Management:

• First-line: fiber, fluids (≥3 extra glasses/day), exercise.
• Docusate 100 mg up to TID; polyethylene glycol (Miralax) daily.
• Avoid chronic stimulant laxatives (senna, bisacodyl, milk of magnesia), which carry a risk of dependence and paradoxical worsening.

Drowsiness / fatigue

Generally mild for an SNRI (far less than a tricyclic), but it happens.

Management: shift to bedtime dosing. If persistent and the drug is otherwise working: modafinil 200–400 mg/day, armodafinil, or a psychostimulant.

Apathy / emotional blunting

The serotonergic "flatness" that patients describe as feeling less: less low, but also less everything. Dose-dependent.

Management: lower the dose (often enough), add bupropion or a stimulant, or switch class if it persists.

Weight

Mild on average, often near weight-neutral, sometimes modest gain over time (serotonergic effects on appetite/leptin). Not a mirtazapine-level problem.

Management: diet/exercise counseling; bupropion augmentation (the one antidepressant associated with weight loss); topiramate 100–150 mg/day (watch cognitive dulling); metformin for the metabolically vulnerable.

Other

• Headache, dizziness: common early, usually transient.
• Hyponatremia / SIADH: class effect, higher risk in the elderly; suspect it with new confusion or lethargy.
• Bleeding risk: serotonergic platelet inhibition raises GI-bleed risk, additive with NSAIDs/anticoagulants (less data than for SSRIs, but treat it as a real class effect).
• Mydriasis / narrow-angle glaucoma: rare, noradrenergic.

Part 6: Toxicity, Overdose, and the Boxed Warning

Hepatotoxicity: the signature safety concern

Practical protocol:

• Screen alcohol use before starting.
• Check baseline ALT and at least once during treatment.
• Avoid in liver disease and heavy drinkers: use desvenlafaxine instead (glucuronidated, minimal hepatic risk, minimal CYP involvement).
• Stop the drug at signs of hepatic injury (rising ALT, RUQ pain, jaundice, dark urine, unexplained fatigue). Don't wait it out.

Overdose

Duloxetine in overdose is far less lethal than a tricyclic. Concerns are serotonergic (serotonin syndrome, especially in combination or co-ingestion), seizures, and, at high doses, hepatotoxicity. Management is supportive; overdose is an ED evaluation. There is no specific antidote.

Serotonin syndrome

Real risk when combined with other serotonergic agents (MAOIs, contraindicated; linezolid; St. John's wort; other antidepressants; dextromethorphan). Duloxetine's own CYP2D6 inhibition can raise co-administered serotonergic drug levels, compounding risk. Note the reassuring update on one old worry: triptans + SNRIs are safe. The historical FDA serotonin-syndrome warning was refuted by large studies (Orlova, JAMA Neurol 2018).

Boxed Warning: Suicidality

Part 7: Drug Interactions

Duloxetine sits in the middle of the SNRI interaction spectrum: cleaner than venlafaxine in some respects, messier than the glucuronidated agents (desvenlafaxine, levomilnacipran).

Duloxetine as a CYP2D6 inhibitor

Duloxetine is a moderate CYP2D6 inhibitor (in the same family as paroxetine and fluoxetine, though generally regarded as less potent than those two). It raises levels of CYP2D6 substrates:

• Tricyclic antidepressants: can rise substantially; if combining, cut the TCA dose and monitor levels (TCAs have cardiac toxicity at high levels).
• Certain antipsychotics (aripiprazole, risperidone, brexpiprazole): raised levels, akathisia/EPS risk.
• Some beta-blockers, stimulants (amphetamines), and other 2D6 substrates: monitor.

Duloxetine's own metabolism (drugs/behaviors that affect its level)

Duloxetine is metabolized by CYP1A2 (primary) and CYP2D6.

• Strong CYP1A2 inhibitors (fluvoxamine, ciprofloxacin, other quinolones) markedly raise duloxetine levels; avoid or dose-reduce. Fluvoxamine is the classic offender.
• Smoking and cannabis induce CYP1A2 and lower duloxetine levels. Clinically meaningful within ~3 days; reverses over ~1 week. A patient who quits smoking (or switches to vaping/nicotine gum) can see duloxetine levels rise; watch for emergent side effects.

The absolute one: MAOIs

• Never combine with an MAOI (serotonin syndrome, potentially fatal).
• 14-day washout in each direction (MAOI ↔ duloxetine).
• 5-week washout when switching from fluoxetine to an MAOI (long half-life of norfluoxetine).

Additive-risk combinations

• Serotonergic agents (other antidepressants, linezolid, St. John's wort, dextromethorphan): serotonin syndrome risk.
• NSAIDs / anticoagulants / antiplatelets: additive GI-bleed risk via platelet serotonin depletion.
• Sympathomimetics (stimulants, decongestants): theoretical additive pressor effect, though duloxetine's own BP effect is minimal.

Part 8: Special Populations

Pregnancy

No dedicated pregnancy category in the source notes, and the drug-specific data are limited, but the class picture is reassuring. A large study (Mezzacappa, JAMA Pediatr 2017) found SSRIs/SNRIs are not associated with congenital malformations or autism. Third-trimester exposure carries a small risk of poor neonatal adaptation (jitteriness, irritability, feeding/tone problems, self-limited) and rarely PPHN.

Framework: untreated depression in pregnancy has real risks of its own; decide case-by-case, minimize total exposures, and, when feasible, favor an agent the patient has previously responded to. Duloxetine is clearly preferable to paroxetine (older Category D, cardiac malformation concern) if the choice is between those two. If starting fresh in pregnancy, many clinicians default to a better-studied SSRI (sertraline), but duloxetine is not contraindicated.

Lactation

Limited data; duloxetine is excreted in breast milk in small amounts (relative infant dose is low). Sertraline is the best-studied and preferred antidepressant in lactation; duloxetine is probably acceptable but less characterized. If used, monitor the infant clinically.

Elderly (65+)

• Start low, go slow. 30 mg or lower to start.
• Effective in geriatric depression and GAD; 60 mg is as good as 120 mg here too.
• No cardiac or anticholinergic toxicity the way a tricyclic has, a real advantage in this population.
• Watch: hyponatremia/SIADH (higher baseline risk in elderly), falls, and the hepatotoxicity issue if polypharmacy or hepatotoxic co-meds are present.
• If medically complex: desvenlafaxine is the cleaner SNRI (fewer interactions, liver-sparing).
• One geriatric study (mean age >70) showed duloxetine improved verbal learning/memory; vortioxetine edged it on processing speed, and both are cognitively benign choices.

Hepatic impairment

Avoid. This is duloxetine's contraindication population. Glucuronidation is impaired in liver disease, and the drug already carries a hepatotoxicity signal. Use desvenlafaxine (glucuronidation-based but the safer choice here) or another class.

Renal impairment

Generally safe, with no major renal metabolism. In severe renal impairment (and ESRD), avoid or use with caution; exposure to the (inactive) metabolites rises. Mild-to-moderate impairment usually needs no adjustment. (Contrast with desvenlafaxine and venlafaxine, which need renal dose reductions.)

Pediatric / adolescent

• FDA-approved for pediatric GAD (ages 7+), a genuine niche.
• For pediatric depression, fluoxetine and escitalopram are the FDA-approved, preferred first-line agents; for pediatric OCD, sertraline. Duloxetine is not first-line for these.
• The boxed suicidality warning applies, as with all antidepressants under 24.

Part 9: Discontinuation, Do This Carefully

This is where duloxetine bites. It has a short half-life (~12 hours), which means plasma levels fall fast when doses are missed or the drug is stopped, and that produces one of the worst discontinuation syndromes in the antidepressant pharmacopeia. Duloxetine sits squarely in the high-risk withdrawal group with venlafaxine, desvenlafaxine, and paroxetine.

How to stop

• Reduce by ~25–50% every 2–4 weeks, slowing further near the end (hyperbolic tapering: the last few milligrams are the hardest).
• The fluoxetine bridge: switch to fluoxetine (long half-life, self-tapering) and then taper the fluoxetine, an established psychopharmacology pearl for tough SNRI discontinuations, though it lacks RCT proof.
• Use the smaller capsule strengths (20, 30 mg) to make fine steps; some patients need to bead-count the delayed-release granules for the final reductions.
• Warn every patient at the start, not at the end: "If you ever want to come off this, we do it slowly and together. Stopping suddenly makes people feel awful, and it's avoidable."

Part 10: Duloxetine vs the Alternatives

vs SSRIs (escitalopram, sertraline, paroxetine). No efficacy edge for depression alone; an SSRI is usually the cheaper, cleaner first choice. Duloxetine differentiates only when you want (a) a pain-syndrome indication, (b) SNRI pharmacology for a partial SSRI responder, or (c) FDA-backed GAD coverage. On sexual dysfunction, duloxetine beats paroxetine and roughly ties escitalopram. On discontinuation, long-half-life fluoxetine is far gentler.

vs Venlafaxine (Effexor XR). The most important intra-class comparison. Efficacy for depression and neuropathic pain is essentially interchangeable. Duloxetine wins on blood pressure (no pressor effect vs venlafaxine's dose-dependent hypertension), on titration (simpler), and modestly on discontinuation severity (bad, but less brutal than Effexor). Venlafaxine wins on not carrying a hepatotoxicity flag and has a clearer dose-response curve you can push. Net: duloxetine for the hypertensive/CV-cautious patient; venlafaxine when you want to reliably dose-escalate and the liver isn't a concern.

vs Desvenlafaxine (Pristiq). Similar efficacy. Desvenlafaxine's real edge is cleanliness: glucuronidated, minimal CYP450, no 2D6 inhibition, liver-sparing, making it the preferred SNRI for polypharmacy, hepatic disease, and medically complex patients. Its liability is a flat dose-response (50 mg ≈ 100 mg; can't meaningfully escalate). Duloxetine gives you dose flexibility (relevant for fibromyalgia) and pain labeling desvenlafaxine lacks.

vs Tricyclics (amitriptyline, nortriptyline). For neuropathic pain and fibromyalgia, TCAs are at least as effective (NNT ≈ 3 vs duloxetine 5.2) and remain rheumatology's first-line for fibromyalgia. Duloxetine wins decisively on safety and tolerability: no anticholinergic burden, no cardiac conduction toxicity, safe in overdose, far better in the elderly. The trade is efficacy-per-milligram (TCA) vs safety and ease (duloxetine).

vs Bupropion. Bupropion is the better choice for uncomplicated depression where you want to avoid sexual dysfunction and weight gain and add some activation. Duloxetine wins when there's anxiety (bupropion can worsen it) or pain (bupropion does nothing for pain). Different tools.

vs Mirtazapine. Mirtazapine for the patient who needs sleep and appetite (and wants to avoid sexual dysfunction); duloxetine for the patient who can't afford the sedation and weight gain, or who needs pain coverage. Both are relatively SD-sparing.

vs Vortioxetine. Vortioxetine has a marginal cognitive edge (a few DSST points, clinically modest, on the order of a cup of coffee) and a slight low-dose sexual advantage that vanishes at therapeutic doses. Duloxetine is generic and cheap with less nausea. In most cases vortioxetine's benefits don't justify its cost and GI burden over duloxetine.

Why is duloxetine reasonably popular? Unlike its SNRI siblings it holds a legitimate footprint (roughly a tenth of the antidepressant market at its peak), driven by the pain indications and the GAD data. Just don't confuse market presence with efficacy superiority: against SSRIs for mood alone, it has none.

Mechanism: The Short Version

Duloxetine inhibits presynaptic reuptake of serotonin and norepinephrine. It has no direct postsynaptic receptor agonism or antagonism, which is why its side-effect profile is "clean" for an older-style antidepressant (no antihistamine sedation, no anticholinergic load beyond mild dry mouth, no alpha-blockade). Despite the "SNRI" label it is ~20:1 selective for serotonin over norepinephrine, more balanced than venlafaxine (120:1) but still functionally a serotonergic drug with a noradrenergic supplement that becomes more relevant at higher doses. The noradrenergic component plausibly underlies its analgesic effect in neuropathic pain (descending pain-inhibitory pathways use both transmitters) and its sweating/constipation/urinary side effects. Metabolism is hepatic via CYP1A2 and CYP2D6, half-life ~12 hours.

The class-level humility, worth remembering: norepinephrine selectivity has never reliably predicted clinical benefit. The promised "noradrenergic" gains in energy, focus, and motivation don't materialize in the data. Treat duloxetine as a serotonergic antidepressant with useful pain labeling and BP neutrality, not as a fundamentally different mechanism from an SSRI.

The Bedside Cheat Sheet