Cymbalta (Duloxetine)

Cymbalta is the drug I reach for when the depression has a physical component...

Sections

1. The patient-autonomy piece, because it should be said up front
2. Dosing
3. The bead-counting taper, which is the thing nobody warns you about
4. Side effects, in plain language, in the order they actually come up
5. What’s nice to hear
6. How it stacks up against the alternatives
7. The liver thing
8. The pattern that ends up on Cymbalta
9. What not to do
10. Who shouldn’t take it
11. Bottom line
12. Sources

Cymbalta is the drug I reach for when the depression has a physical component… when guys come in saying their back hurts, their joints hurt, they’re tired in a way sleep doesn’t fix, and yeah they’re also depressed but it’s not just sad, the body is checked out too. SSRIs don’t do much for that. SNRIs sometimes do, and Cymbalta is the SNRI I reach for the most.

It’s a serotonin-norepinephrine reuptake inhibitor, which means it does the SSRI thing on top of working on norepinephrine, which is the neurotransmitter that runs your get-up-and-do-stuff system and, weirdly, also the wiring your spinal cord uses to dial pain signals up or down before they reach your brain. That second piece is why duloxetine is FDA-approved for fibromyalgia, diabetic neuropathy, and chronic musculoskeletal pain, on top of depression and generalized anxiety. It’s basically a Swiss army knife of a drug if the diagnosis fits.

The patient-autonomy piece, because it should be said up front

Before we get into the dosing: if you want Cymbalta, you get Cymbalta. I’m a provider, not a parent. My job is to lay out the honest take on what it’ll do and what the trade-offs are, your job is to decide. Sometimes I’m writing a Cymbalta script I’d personally have voted against, which is fine, the appointment isn’t mine. The most I’ll do is make it a disapproving yes when I have reservations, where you walk out with the prescription and a clear sense of what I’d watch for and why I wasn’t thrilled. I hardly ever say no. With Cymbalta the main reservation is the discontinuation profile, which is real and worth knowing about before you start, not after.

Dosing

Start at 30 mg for a week or two, then 60 mg, which is where most adults end up. The data above 60 mg is unimpressive, you mostly trade modest extra benefit for noticeably more side effects. I rarely go above 90 mg. If 60 mg isn’t working after eight weeks at that dose, the answer is usually to switch or augment, not push higher.

It comes in a capsule full of little enteric-coated beads, which becomes relevant when we taper, because the discontinuation on this drug is rough and you have to taper slowly.

The bead-counting taper, which is the thing nobody warns you about

Coming off Cymbalta cold turkey or even on the manufacturer’s recommended taper schedule is awful for a lot of people. Brain-zaps, dizziness, flu-feeling, mood crash, the works. The reason is the half-life is short (around twelve hours, which means the drug clears out fast) and the dose increments the manufacturer makes (20, 30, 60 mg) are too coarse for how sensitive most people’s systems are to dropping off.

So we do the bead taper. You open the capsule, count the beads, remove a few each week. Some people do ten percent every two weeks, some go slower. A 60 mg capsule has somewhere between 250 and 350 beads depending on the manufacturer, and yes, this is annoying, and yes, this is how a lot of guys get off the drug without two months of feeling like they have the flu. The pharmacy is not going to help you with this and your prescriber may not even know about it. Now you do.

Side effects, in plain language, in the order they actually come up

Nausea in the first two weeks, more than with Lexapro, take it with food and it mostly goes away by week three… dry mouth and sometimes constipation, tolerable for most guys, annoying for some… sexual side effects, slightly less than the SSRIs in most studies but not zero, maybe twenty to thirty percent of guys notice something (same fix as with Lexapro, add Wellbutrin or switch)… sweating, this one is weird and it catches guys off guard, norepinephrine drives sweating and Cymbalta cranks up norepinephrine, so a not-small number of patients end up sweating through their shirts in meetings for no obvious reason, sometimes it eases in a few months, sometimes it doesn’t… blood pressure can creep up a little, not usually clinically significant but worth checking at the three-month visit especially if you were borderline hypertensive going in… some guys feel out of it on it, like their edges are softer than usual, and for those guys we either drop the dose or move on.

What’s nice to hear

This is the part the side-effects-heavy version of this conversation tends to skip. For the patient where the diagnosis actually fits, Cymbalta is genuinely useful, in a way SSRIs aren’t. The chronic back pain that’s been a six out of ten for years drops to a three or four. The bone-deep-tired version of fatigue starts to lift. The patient who’d been describing himself as “checked out” stops feeling that way. The drug has FDA-approved data for chronic pain conditions that SSRIs don’t have, and the clinical experience matches the data. Two months in, the guy who was on a low-dose opioid he wanted off of, and was drinking more than he should, is sometimes able to stop drinking on his own once the mood and the pain are both improving, without anybody lecturing him. That’s a real outcome and it’s worth saying out loud.

How it stacks up against the alternatives

Effexor (venlafaxine) is the other big SNRI and the comparison comes up a lot. Effexor at low doses (75 to 150 mg) acts mostly like an SSRI because the norepinephrine effect only really kicks in above 150 mg. Cymbalta hits both serotonin and norepinephrine across its dose range, which means the SNRI properties are present even at 30 or 60 mg. For pain indications this matters… Cymbalta has FDA approval for chronic pain conditions and Effexor doesn’t. For pure depression with no physical component, Effexor and Cymbalta are roughly equivalent in effect, and the choice comes down to side effect profile and discontinuation, both of which are slightly worse on Effexor in most clinical experience. Another tick in the Cymbalta column for the right patient.

Compared to SSRIs in patients with significant physical symptoms (back pain, fibromyalgia, neuropathy, the general bone-deep-tired flavor of depression), Cymbalta tends to do more work. The norepinephrine component is the part that matters for the body symptoms, and SSRIs just don’t reach that wiring.

The liver thing

Cymbalta is metabolized by the liver and there’s a real signal of liver injury, mostly in people who already have liver problems or who drink heavily. The FDA put a warning on it. If you’re a heavy drinker, this is not your drug, full stop, stacking a hepatotoxin (a substance that’s hard on the liver) on a liver that’s already working overtime is the kind of decision future you is going to call current you a dick about. If you drink moderately, fine, but I’d rather not pile hepatotoxic stuff on a liver that’s already metabolizing alcohol every night. Get an LFT panel (liver function tests, the standard blood draw that tells us if the liver is annoyed) at baseline and at three months if you’re in any kind of risk group.

And on the booze side… don’t drink heavily on it, not because there’s some catastrophic interaction lurking in week one, just because alcohol is a depressant and a hepatotoxin and you’re on a depression medication that also stresses the liver, the math doesn’t really pencil out, you can probably figure out where I’m going with this.

The pattern that ends up on Cymbalta

The kind of guy who ends up on this drug is the one whose wife dragged him in, who’d had a back surgery a couple years prior that didn’t fully fix things, who was on a low-dose opioid he wanted to get off of, who was drinking more than he should have been, and who was depressed without really using that word. The way it comes out is usually some version of, “I’m done. Done with work, done with the pain, done.”

Say you start that guy on Cymbalta at 30 mg, go up to 60 mg by week four. Two things tend to happen. The chronic back pain drops a couple of points on the 10-point scale, which the drug actually has data for. And the mood lifts enough that the drinking eases on its own, without anybody lecturing about it, which is usually how that has to happen anyway. The opioid comes off over six months. Two years in, the patient describes the version of himself before this drug as a guy who was checked out of his own life. He’s not anymore.

If you’re a heavy drinker, this is not your drug, full stop… stacking a hepatotoxin on a liver that’s already working overtime is the kind of decision future you is going to call current you a dick about.

What not to do

Don’t skip doses. The half-life is short enough that even a missed day can give you a preview of what the discontinuation will feel like.

Don’t taper on the manufacturer’s schedule unless you’re one of the lucky few who don’t get withdrawal. Do the bead taper, your prescriber will think it’s weird, do it anyway.

Don’t take it if you have uncontrolled narrow-angle glaucoma (a specific eye-pressure condition where this drug class can make the pressure worse). Don’t combine with MAOIs (the older antidepressants with the special diet, this should be obvious but it has to be said).

If you’ve got a real history of liver problems, pick a different drug.

Who shouldn’t take it

People with significant liver disease. People on MAOIs (the usual rules apply). People with uncontrolled narrow-angle glaucoma. People who already have uncontrolled high blood pressure, because the norepinephrine effect will nudge BP up another notch and you don’t want to start at 145/95.

Heavy drinkers, again. I know I said it already. I’m saying it again because guys who drink heavily are usually the same guys whose depression has a strong physical component, and who would otherwise be great candidates for Cymbalta, and the decision tree there is to address the drinking first, get an LFT panel, and revisit. Sometimes the depression lifts on its own once the drinking is gone, and we don’t need Cymbalta at all.

Bottom line

Cymbalta is great when the diagnosis fits, which is depression-plus-pain, or depression-plus-the-body-feels-broken, or generalized anxiety with physical symptoms. It’s a hassle to get off of and you need to know that going in. For the right patient it’s one of the most useful drugs in the cabinet. For the wrong one, it’s a Lexapro that gives you sweats and a worse taper. The decision to go on it, or come off it later, is yours. My job is the honest take, your job is the choice.

Sources

1. U.S. Food and Drug Administration. Cymbalta (duloxetine hydrochloride) Prescribing Information. NDA 021427. FDA; 2023. FDA label.
2. Siddiqui F, Petersen JJ, Juul S, et al. Beneficial and harmful effects of duloxetine versus placebo, ‘active placebo’ or no intervention for adults with major depressive disorder: a systematic review with meta-analysis and trial sequential analysis of randomised clinical trials. BMJ Open. 2025;15(2):e082853. PMID 39920066.
3. Cipriani A, Furukawa TA, Salanti G, et al. Comparative efficacy and acceptability of 21 antidepressant drugs for the acute treatment of adults with major depressive disorder: a systematic review and network meta-analysis. Lancet. 2018;391(10128):1357, 1366. PMID 29477251.