I started Prozac 40 mg on Friday and felt lighter by Sunday. Is that the drug?
The clock is weeks, not a breakfast lift
Forty milligrams does not lift a Monday the way caffeine does. If a reader came here for a same-day mood story, they have the wrong folio. The US label says the full antidepressant effect may be delayed until four weeks of treatment or longer. That sentence is older than most threads, and it is still the one I write at the top of an onset file.
This shelf locks Prozac talk to 40 mg because that is the later rung people argue about - the capsule they imagine will finally 'kick in' after 20 mg felt quiet. The label's own order is the opposite of a loading-dose fantasy: start 20 mg in the morning, wait several weeks, then consider a rise. 40 mg is that rise. It is not a first swallow for most adults with major depression.
I will not rebuild a dosing monograph. OCD, bulimia, panic, and the olanzapine combination have their own clocks and ceilings. You can open those on the fluoxetine slip. Here the only job is time: when a lift is a fair question, why week one lies, why norfluoxetine outlasts the blister, and why 'long half-life means no discontinuation' is a myth wearing a pharmacokinetic coat.
Norfluoxetine outlives the parent capsule
Norfluoxetine is why a missed capsule rarely feels like a cliff. CYP2D6 turns fluoxetine into an active metabolite whose elimination half-life after acute or chronic use is 4-16 days. Parent fluoxetine itself stretches from 1-3 days after a first dose to 4-6 days in chronic use, because the drug inhibits its own metabolism. Together they accumulate. Together they leave slowly.
That persistence is the onset file's double edge. On the way in, you cannot judge a 40 mg step by Friday. On the way out, active drug is still on the shelf in plasma for weeks. The label warns that this matters when you stop, and when you start a drug that will fight with fluoxetine or norfluoxetine after the last Prozac date on the blister. The classic example is an MAOI: the wait after fluoxetine is five weeks, not the shorter gap used after many other SSRIs.
About 7% of people are CYP2D6 poor metabolizers. Their parent half-life can more than double. I do not genotype everyone on this desk. I do remember that a 'standard' four-week clock is already a mean, and some bodies sit to the right of it. Side effects that bloom late, and benefits that arrive late, are both more plausible on this molecule than on a one-day SSRI.
Weekly 90 mg fluoxetine was once marketed because the combined half-life can cover a gap. Peak-to-trough swing is wider on a weekly capsule than on a daily 20 mg. I mention it only as proof that the manufacturer knew the persistence was real. It is not a suggestion to invent your own every-other-day scheme from a 40 mg blister.
What a four-week file asks the desk
The desk question after week four is not 'did breakfast feel different.' It is whether sleep, interest, guilt, energy, and suicidality have moved on a scale you can repeat. STAR*D used QIDS at visits and a manual for when to raise the dose. You do not need their exact forms. You need some form. A vibe is how people stay on 40 mg for a year without ever knowing if the rung did work.
If week four on 20 mg is a flat line and the capsule is tolerated, the label's door to 40 mg is open. If week four on 40 mg is a flat line, the next conversation is about diagnosis, adherence, alcohol, thyroid, and a different class - not about a secret third Prozac strength that will finally act like caffeine. STAR*D Level 2 existed because Level 1 left most people unremitted. That is ordinary, not a personal failure.
If week one is agitation, insomnia, or a new suicidal spike, that is not an onset-file delay. That is a safety visit. Especially under 25. I would rather a reader stop and call than 'tough out' a boxed warning because this page said wait four weeks. The four-week clock is for efficacy in someone who is holding together. It is not a gag on an emergency.
Partners notice the clock too. They see the first days of restlessness and decide the drug is 'making them worse,' or they see no change and decide it is 'fake.' Hand them the STAR*D means - 5.7 and 6.7 weeks - and the label's four-week line. It will not make anyone patient. It will at least stop the household from running a 48-hour trial.
Week-one mood stories fail a shelf-check
Why a week-one mood story fails the shelf-check
- Steady state for fluoxetine plus norfluoxetine sits around 4-5 weeks, not a Monday morning.
- STAR*D mean response took 5.7 weeks on another SSRI in real clinics - the clock is class-shaped, not brand-magic.
- Early activating effects (insomnia, restlessness) get misread as a lift, then crash when the first week ends.
- Dose changes, including a step to 40 mg, are not fully reflected in plasma for several weeks - the label says so in plain words.
A same-day brightness is usually something else - sleep, agitation, a placebo week. Fluoxetine can feel activating early. That is a real experience and a real reason some people stop. It is not the depression endpoint used in registration trials or in STAR*D. Those files asked for weeks of dosing and a scored drop, not a Tuesday afternoon text that says 'I think it's working.'
Classic SSRI teaching, which STAR*D then made painfully concrete, is that a fair efficacy look sits at week four and often later. People who remit after week eight are not outliers in that Level 1 set. They are a slice of the success column. If a clinic books the only review at day ten, it has built a machine for false failures.
There is a kinder early signal worth naming without turning it into a same-day claim. Some people sleep a little better, or cry with a shorter fuse, in week two. Those crumbs can predict who will later respond. They are not a green light to raise to 40 mg on impulse, and they are not a red light to stop if they are missing. The shelf-check is: early crumbs optional, scored review mandatory, same-day lift off the card.
Younger patients need a darker early watch that has nothing to do with efficacy. The boxed warning on suicidal thinking and behavior in children, adolescents, and young adults is a first-weeks problem as much as a later one. A new burst of agitation or hopelessness is a same-week reason to call, not a reason to 'give it time' the way we give a HAM-D time. Do not confuse those two clocks.
Forty milligrams arrives after the first card
The US label starts adults at 20 mg in the morning. Studies that compared 20, 40, and 60 mg with placebo found 20 mg enough for a satisfactory antidepressant response in most adults. A rise is a second decision, after several weeks, if the first card is not enough. Doses above 20 mg may be once daily or split morning and noon. The ceiling is 80 mg per day. That is the ladder. 40 mg is a middle step on it, not the opening move.
Because parent plus metabolite take weeks to settle, a jump to 40 mg on day three is not 'being decisive.' It is changing a dose before the first dose has been given a plasma chance. The label's long-half-life warning is explicit: changes will not be fully reflected in plasma for several weeks. That sentence governs titration and withdrawal. It is why this review refuses to become a second dosing monograph with a week-by-week recipe. The recipe is wait, then review, then maybe rise.
After 30 days at 40 mg daily, the label quotes fluoxetine levels of 91-302 ng/mL and norfluoxetine of 72-258 ng/mL. Steady-state pictures after prolonged dosing look like the picture at 4-5 weeks. If a reader feels nothing on day five of 40 mg, they have described expected pharmacokinetics, not a failed drug. If they feel wired on day five, that can be an activating effect. It is still not the HAM-D endpoint.
OCD is the one labelled clock I will mention because people mix it with depression. The OCD line says the full therapeutic effect may wait until five weeks or longer, and the usual range runs 20-60 mg. That is a longer, often higher file. It does not pull depression's four-week sentence down to a weekend.
STAR*D timed response in real clinics
Day 0-3
Parent fluoxetine half-life after a first capsule: about 1-3 days. A mood lift is not the endpoint.
Week 1
Sleep, jitter, gut change, or a placebo week can look like 'it worked.' The label does not stamp efficacy here.
Week 4
US Prozac label: full antidepressant effect may be delayed until 4 weeks or longer. This is the first honest review point.
After several weeks
Label door to a later rung: consider a rise above 20 mg if the first card is not enough. 40 mg is that later card, not breakfast.
Week 5.7-6.7
STAR*D Level 1 (citalopram, not fluoxetine): mean time to response 5.7 weeks, to remission 6.7 weeks.
Week 8+
A substantial share of STAR*D remitters crossed the line at or after week 8. Stopping at week 2 because 'nothing happened' misreads the file.
Days 4-16 after a stop
Norfluoxetine half-life 4-16 days. Active drug persists for weeks. That eases some cliffs. It does not erase discontinuation.
STAR*D did not study fluoxetine as the first tablet. Level 1 used citalopram in 2,876 outpatients across 23 psychiatric and 18 primary-care sites, flexible dose, up to 14 weeks, measurement-based visits. Rush, Trivedi, and colleagues published the Level 1 file in the American Journal of Psychiatry in 2006. I still teach it on an onset desk because it is the largest ordinary-clinic clock we have for an SSRI, not because the molecule is the same.
Remission - the primary goal, not a 50% dip - was 28% on the Hamilton 17-item scale (score 7 or less) and 33% on QIDS-SR (score 5 or less). Response, a 50% QIDS-SR drop, was 47%. Mean exit dose sat at 41.8 mg per day. Mean time to response was 5.7 weeks. Mean time to remission was 6.7 weeks. A substantial share of people who eventually remitted did so at or after week 8. That last line is the one week-one culture never quotes.
Nearly 80% of the analysed sample had chronic or recurrent depression. Comorbidity was the rule, not the clean Phase 3 volunteer. Primary-care and psychiatry remission rates did not split. The lesson I bind is blunt: even with a coordinator, a manual, and dose increases, an SSRI clock in the real world is counted in weeks, and a fair slice of the wins arrive after the date a frustrated person has already stopped.
Note what STAR*D is not. It is not a fluoxetine-versus-placebo registration trial. It is not a licence to wait forever without a plan. People who could not tolerate the drug, or who stayed severely ill at a tolerated ceiling, moved to Level 2. The onset file borrows the timing and the honesty about remission rates. It does not borrow a molecule.
Long washout is not the same as no withdrawal
People hear 'long half-life' and assume they can stop cold without a flicker. The pharmacokinetics do blunt the cliff that short-acting agents - paroxetine, venlafaxine - are famous for. A slow decay of norfluoxetine is a built-in taper for some bodies. It is not a guarantee. Discontinuation symptoms (dizziness, electric-shock sensations, irritability, flu-like days, vivid dreams) still happen on fluoxetine, especially after high dose or long use, and especially in people who are sensitive to any drop.
The label's own logic cuts both ways. Because plasma does not finish changing for weeks, a stop is not 'clean' on Friday, and a new interacting drug is not safe the moment the blister is empty. Clinicians sometimes switch a short SSRI to fluoxetine to ease a brutal withdrawal. That trick uses persistence. It does not prove that persistence equals zero symptoms. Some authors dislike the switch tactic. I treat it as a specialist tool, not a home recipe.
The myth I want off this shelf is the internet version: 'Prozac has no discontinuation syndrome, so you can vanish.' Vanish and you may feel fine. Vanish and you may feel wretched two weeks later, when you have already told yourself the drug 'never did anything.' Vanish and start an MAOI too soon and you have a safety problem, not a comfort problem. Bind the long clock as a caution, not as a dare.
Bind the onset file and stay in the room
Bind the onset file as a waiting slip, not a same-day stamp. 20 mg is the usual first card. 40 mg is a later rung after several weeks. The label's antidepressant clock starts at four weeks and does not apologise. STAR*D, on a sister SSRI, put mean response at 5.7 weeks and mean remission at 6.7, with a slice of the remissions after week eight. Norfluoxetine keeps the lights on after the last capsule - useful, and a reason MAOI waits run to five weeks.
A week-one glow is not the trial endpoint. A long half-life is not a free pass to disappear. Young people still need the suicide watch in the first weeks. None of that is a checkout. It is a desk that stays in the room while the metabolite accumulates.
Read the fluoxetine folio for the rest of the molecule. If you want a completely different evidence shape - counted chips and named parasite platforms - open Stromectol field ink. Then take the clock to your own clinician before anyone moves a milligram.
Last Updated
Reader mail
Reader questions on this review
Answered by Dr. Ingrid Halvorsen, MD · Internal medicine & clinical pharmacology
Onset questions arrive as week-one disappointments and 40 mg daydreams. I answer with the label delay and the STAR*D clock, not with a pep talk. General teaching. I have not met you.
I would not stamp it as the antidepressant effect. The US label says a full effect may wait until four weeks or longer, and after 30 days at 40 mg the parent and norfluoxetine levels are still the ones the manufacturer quotes as a chronic picture. A weekend lift is more often sleep catching up, a crisis easing, caffeine, or an activating burst that can flip into jitter. Enjoy a better Sunday if you have one. Book the real review at week four, and call sooner if the 'lift' is agitation or a new dark turn. That early watch matters more under 25.
STAR*D used citalopram. Why do you keep putting it on a fluoxetine page?
Because I need a large, messy, ordinary-clinic clock, and Level 1 is the one we have. 2,876 people, mixed primary care and psychiatry, mean exit dose 41.8 mg, remission 28% on HAM-D and 33% on QIDS-SR, response 47%, mean 5.7 weeks to response and 6.7 to remission, with a chunk of remitters at or after week 8. That is how long an SSRI job takes when the sample looks like a waiting room. I say the molecule out loud every time so nobody thinks STAR*D was a Prozac registration trial. The cousin fact on this slip is the Prozac sentence: wait at least four weeks before you call the first card a failure.
My clinician jumped me from 20 mg to 40 mg after ten days because I 'needed it to work faster.' Fair?
The label's order is start 20 mg, consider a rise after several weeks if improvement is insufficient. Ten days is not several weeks, and plasma will not have finished reflecting the first dose. I am not in the room - there are cases (severe illness, prior known dose, a specialist plan) where a faster rise is argued. As default teaching, a ten-day jump is impatience wearing a milligram. It also raises the chance you will blame 40 mg for side effects that were still the 20 mg accumulating. Ask for the reason on paper. If the reason is only 'so you feel it,' that is not an onset-file reason.
If norfluoxetine lasts up to 16 days, can I skip capsules on weekends to 'reset'?
No. Weekend holidays are how people get irregular plasma, irregular sleep, and a false story that the drug is 'unpredictable.' Chronic fluoxetine already has a 4-6 day parent half-life; norfluoxetine adds 4-16 days. That persistence covers an accidental missed day better than a short SSRI would. It does not make an invented every-other-day schedule from a 40 mg blister into a therapy. The old weekly 90 mg product was a specific capsule with its own peak-to-trough data, not a licence to improvise. Take the prescribed rhythm. Tell your clinician if swallowing every day is the actual problem.
I stopped 40 mg two weeks ago and feel electric zaps. I thought Prozac had no withdrawal.
Long washout lowers the odds of a brutal cliff. It does not set the odds to zero. Discontinuation symptoms are well described on fluoxetine: dizziness, shock-like sensations, irritability, flu-ish days, vivid dreams. They can arrive late, because norfluoxetine is still falling in week two and three. The myth that 'Prozac has no discontinuation syndrome' is a half-life slogan, not a trial result. Contact the prescriber who stopped it - sometimes a slower taper, a brief restart, or a structured switch is the kinder path. Do not start an MAOI in that window; the labelled wait after fluoxetine is five weeks.
Does a better week-two sleep mean I will remit, or should I still wait for week four?
A crumb in week two - slightly better sleep, a shorter crying jag - can be a hopeful early signal. It is not remission, and STAR*D's mean remission sat at 6.7 weeks with more people crossing later. Keep the week-four review. Keep a scored list so you are not relying on a partner's mood to judge yours. If week two is worse - new agitation, hopelessness, especially if you are young - that crumb rule does not apply. That is a same-week call, not a 'wait for the metabolite' speech.
Why is the MAOI gap five weeks after fluoxetine when other SSRIs use two?
Because parent plus norfluoxetine are still sitting in plasma long after the last capsule. The label's long-elimination section exists for exactly this: dose changes and interacting drugs are not finished when the blister is empty. An MAOI plus leftover fluoxetine is a serotonin-syndrome setup. Two weeks is a common gap after shorter agents. Five weeks after fluoxetine is the conservative, labelled wait. Do not shorten it because you 'feel washed out already.' Feeling empty is not the same as a clear CYP2D6 file.
If most STAR*D patients did not remit on the first SSRI, is starting fluoxetine even worth it?
Remission in Level 1 was about a third. Response was about half. That is sobering and still higher than 'do nothing' in people who are ill enough to be in that study. The point of an onset file is not to sell a miracle rate. It is to stop people from abandoning a tolerable first card at day nine, and to stop them from treating 40 mg as a same-day rescue. Many who remitted needed eight or more weeks and a measured dose. Many who did not remitted later on a different plan. Worth is a conversation with your own clinician, with severity and support on the table - not a social-media remission percentage.
Where is the rest of the Prozac file if this page will not list every dose?
On the fluoxetine folio slip: indications, the boxed warning, interactions, and the labelled rungs this review refused to turn into a second monograph. For a different evidence shape entirely - four 3 mg chips and named parasite platforms - see Stromectol field ink. Bring the four-week clock, your other medicines, and any young-person suicide watch to your own prescriber before a milligram moves.
These replies are teaching, not a treatment plan for the person who asked. Take your own history, labs and medicine list to a prescriber who can see all three.