the public record, side by side

Lexapro (escitalopram) vs Prozac (fluoxetine)

Lexapro and Prozac, compared on what the public record actually supports: side effects by rate from each FDA label, what the largest trial comparison found, who takes each one, who pays, what the labels warn, and what the litigation record holds. Sources on every number.

the short answer

is Lexapro or Prozac more effective? in the largest comparison of antidepressant trials (522 trials, 116,477 people, Lancet 2018), both beat placebo on response at about 8 weeks: escitalopram odds ratio 1.68 (95% CrI 1.50 to 1.87), fluoxetine 1.52 (1.40 to 1.66). the authors called the differences between drugs mostly modest, with wide intervals. the 2009 Cochrane review pooled the three trials with response data and found no difference in response at 6 to 12 weeks (odds ratio 0.81, 95% CI 0.60 to 1.10) or remission (0.86, 0.65 to 1.15), with a small edge for escitalopram on the depression-score change; the labels differ mostly in kinetics: fluoxetine’s active metabolite lasts 4 to 16 days, so its label requires 5 weeks before an MAOI against 14 days for escitalopram, and fluoxetine’s warnings section carries a QT prolongation item that escitalopram’s does not.
which has fewer side effects? the labels say their trials “cannot be directly compared”, because each drug was tested in different people under different rules. what each label prints: Lexapro (escitalopram): 6% stopped for an adverse reaction vs 2% on placebo (Lexapro label, Revised:4/2024 (as printed in the SPL)). Prozac (fluoxetine): the label prints no discontinuation rate in the section we transcribed. the reactions both labels list, side by side, are in the table below. in the 2018 network, all-cause dropout against placebo was 0.90 for escitalopram and 0.88 for fluoxetine.
which should i take? not a question this page answers. the odds ratios are averages across thousands of strangers; your history, other medications and what you can live with are not in them. the questions for your prescriber below are built from what is on this page.

What this page will not do is rank them. For most same-class pairs, head-to-head efficacy differences are smaller than the rankings look — the largest network meta-analysis ever run (21 antidepressants, 522 trials) found every drug beat placebo and the gaps between drugs mostly modest and uncertain. Which one fits you turns on your history, your other medications, and side-effect trade-offs — a prescriber conversation. This page arms that conversation; here are the questions worth taking into it →

what the 522-trial comparison found for escitalopram and fluoxetine

the 2018 Lancet network meta-analysis pooled 522 trials comprising 116 477 participants, the median duration of the acute treatment was 8 weeks (iqr 6–8), and compared 21 antidepressants against placebo and against each other on two outcomes: response (at least a 50% fall on the depression scale) and all-cause dropout, its stand-in for acceptability. odds ratios are against placebo. an odds ratio above 1 on response favours the drug; below 1 on dropout means fewer people stopped than on placebo.

drug	response vs placebo, OR (95% CrI)	dropout vs placebo, OR (95% CrI)
Lexapro (escitalopram)	1.68 (1.50 to 1.87)	0.90 (0.80 to 1.02)
Prozac (fluoxetine)	1.52 (1.40 to 1.66)	0.88 (0.80 to 0.96), fewer dropouts than placebo

escitalopram: among the drugs that were more effective than other antidepressants in head-to-head trials; among the more tolerable (fewer dropouts) in head-to-head trials.
fluoxetine: among the least efficacious in head-to-head trials; among the more tolerable (fewer dropouts) in head-to-head trials.

the other direction, in the authors' words. “the certainty of evidence was moderate to very low”; “The random-effects summary SMD for all antidepressants was 0·30 (95% CrI 0·26–0·34)”, a modest average effect; “positive results need to be taken with great caution, because the trials in this network meta-analysis were of short duration”. 46 (9%) of 522 trials were rated as high risk of bias, 380 (73%) trials as moderate, and 96 (18%) as low. 409 (78%) of 522 studies were funded by pharmaceutical companies. the trials excluded people with psychotic, treatment-resistant or bipolar depression, and the paper says it cannot guide the next choice after a first drug fails. a difference between two odds ratios in the table above is not a difference between the drugs unless the direct comparison shows one.

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; doi:10.1016/S0140-6736(17)32802-7. funding of the analysis: National Institute for Health Research Oxford Health Biomedical Research Centre and the Japan Society for the Promotion of Science. The funder of this study had no role in study design, data collection, data analysis, data interpretation, writing of the report, or in the decision to submit for publication. read 2026-10-11 at PMC.

the trials that compared these two directly, and the two label differences that matter day to day

who paid: the Cochrane review is academic (University of Verona and collaborators). of the four trials it pooled against fluoxetine, the review records three (Kasper 2005, Kennedy 2005, SCT-MD-09) as funded by the escitalopram manufacturer, and two of those exist only as unpublished data; the fourth, Mao 2008, ran in six Chinese hospitals and its funder is not stated in the review or the abstract. the review’s own closing line: "the potential for overestimation of treatment effect due to sponsorship bias should also be borne in mind."

the 2009 Cochrane review of escitalopram (searched to July 2008) found four randomised trials against fluoxetine: Kasper 2005 (517 patients aged 65 and over, escitalopram 10 mg vs fluoxetine 20 mg vs placebo, 8 weeks), Mao 2008 (240 patients in China, 10 mg vs 20 mg, 8 weeks), Kennedy 2005 (205 outpatients, 10 to 20 mg vs 20 to 40 mg, 8 weeks, unpublished) and SCT-MD-09 (30 patients, a 5-week sleep study that did not measure efficacy).
pooled response at 6 to 12 weeks: "no evidence that escitalopram was more or less efficacious than fluoxetine" (OR 0.81, 95% CI 0.60 to 1.10; 3 studies, 783 participants; odds ratios written so that below 1 favours escitalopram). remission: OR 0.86 (0.65 to 1.15). on the continuous depression score, escitalopram came out ahead by a small margin (standardised mean difference −0.17, −0.32 to −0.03; 3 studies, 759 participants), which is the "efficacy (citalopram and fluoxetine)" sentence in the review’s conclusions.
acceptability and tolerability, pooled: stopping for any cause OR 0.89 (0.51 to 1.55), stopping for side effects 0.75 (0.44 to 1.28), at least one side effect 0.80 (0.59 to 1.07): none of the intervals excludes 1. excluding the trial with a dropout rate above 20% did not change the result.
the direct trials, as their own abstracts report them. Kasper 2005, in patients with a mean age of 75: "formally, this was a ‘failed study’", neither drug beat placebo on the primary measure; fluoxetine "showed significantly lower efficacy than both escitalopram and placebo"; withdrawals for adverse events were 2.8% on placebo, 9.8% on escitalopram and 12.2% on fluoxetine. Mao 2008: the week-8 HAM-D-17 figures in the abstract are 15.8 for escitalopram and 14.7 for fluoxetine, not significantly different, with "no significant differences in response rates at all visits"; a post hoc analysis favoured escitalopram on two single items, and adverse events "were comparable".
the other side of the ledger: the 2013 Cochrane review of fluoxetine (Magni 2013; 171 studies, 24,868 participants) found fluoxetine less effective than sertraline (OR 1.37, 1.08 to 1.74), venlafaxine, mirtazapine and dosulepin, better tolerated than the tricyclics, and wrote that "the clinical meaning of these differences is uncertain" because most trials failed to report their methods and most were company sponsored. its abstract names no escitalopram result.
the labels, read on 2026-10-21 (Lexapro, DailyMed set 13bb8267, version 45; fluoxetine capsules, set 23899405, version 16, effective 20 August 2026). dose: escitalopram’s maximum is 20 mg a day ("No additional benefits were seen at 20 mg once daily"); fluoxetine’s is 80 mg a day. kinetics: escitalopram’s terminal half-life is "about 27-32 hours"; fluoxetine’s is 4 to 6 days after repeated dosing and its active metabolite norfluoxetine’s is 4 to 16 days, so the fluoxetine label carries a warning section of its own (5.14) saying "changes in dose will not be fully reflected in plasma for several weeks" and notes that "the full therapeutic effect may be delayed until 5 weeks of treatment or longer". the practical consequence is in the contraindications: fluoxetine must not be followed by an MAOI for 5 weeks, escitalopram for 14 days.
QT: the fluoxetine label has a QT prolongation item in its warnings section (5.11, post-marketing reports of QT prolongation and torsades de pointes, "use with caution" in people with risk factors) and contraindicates thioridazine and pimozide; the escitalopram label contraindicates pimozide but carries no QT item among its eleven warnings, reporting instead that in its thorough QT study "none of the patients in the Lexapro group had a QTcF interval >500 msec or a prolongation >60 msec".

limits: three trials with response data, two of them fixed-dose at 10 mg against 20 mg, so "no difference" means "at those doses"; the largest published trial was in people over 65 and failed against placebo; the review is from 2009 and its own authors flag sponsorship bias. the 2018 Lancet network, which pooled far more trials, gives both drugs overlapping intervals against placebo (escitalopram 1.68, 1.50 to 1.87; fluoxetine 1.52, 1.40 to 1.66) and in its head-to-head analysis names escitalopram among the more effective drugs and fluoxetine among the least efficacious, with both among the more tolerable. a long half-life cuts both ways: it is why a fluoxetine dose change takes weeks to show, and both labels carry the same discontinuation-reactions warning and the same advice to taper rather than stop abruptly. none of this is a verdict about you.

Cipriani A, Santilli C, Furukawa TA, et al. Escitalopram versus other antidepressive agents for depression. Cochrane Database of Systematic Reviews 2009; doi: 10.1002/14651858.CD006532.pub2. read 2026-10-21.

side effects by rate: what each label prints, not a head-to-head trial

the labels' own warning, first: “Because clinical studies are conducted under widely varying conditions, adverse reaction rates observed in the clinical studies of a drug cannot be directly compared to rates in the clinical studies of another drug and may not reflect the rates observed in practice.” the two columns below come from different trials, in different people, under different inclusion rules. read each against its own placebo column, not against the other drug.

reaction	Lexapro (N=715) vs Placebo (N=592)	PROZAC (N=2869) vs Placebo (N=1673)
nausea	15% vs 7%	22% vs 9%
dry mouth	6% vs 5%	9% vs 6%
diarrhea	8% vs 5%	11% vs 7%
constipation	3% vs 1%	5% vs 4%
insomnia	9% vs 4%	19% vs 10%
somnolence	6% vs 2%	12% vs 5%
dizziness	5% vs 3%	9% vs 6%
sweating	5% vs 2%	7% vs 3%
decreased appetite	3% vs 1%	10% vs 3%
libido decreased	3% vs 1%	4% vs 1%

Lexapro (escitalopram): 6% stopped for an adverse reaction vs 2% on placebo (Lexapro label, Revised:4/2024 (as printed in the SPL)).
Prozac (fluoxetine): the label prints no discontinuation rate in the section we transcribed.

what makes each table, in the labels' words

Lexapro (escitalopram), Revised:4/2024 (as printed in the SPL), 6 ADVERSE REACTIONS / 6.1 Clinical Trials Experience. inclusion rule: “Reactions included are those occurring in 2% or more of patients treated with Lexapro and for which the incidence in patients treated with Lexapro was greater than the incidence in placebo-treated patients.” trials: “Adverse reactions information for Lexapro was collected from 715 patients with major depressive disorder who were exposed to escitalopram and from 592 patients who were exposed to placebo in double-blind, placebo-controlled trials. Doses ranged from 10 to 20 mg/day; incidences are rounded to the nearest percent by the label itself.” DailyMed
Prozac (fluoxetine), SPL effective 2026-01-08 (the PI carries no printed 'Revised:' footer; only the attached Medication Guide is dated, revised August, 2023), 6 ADVERSE REACTIONS / 6.1 Clinical Trials Experience. inclusion rule: “...treatment-emergent adverse reactions that occurred in 2% or more patients treated with PROZAC and with incidence greater than placebo who participated in US Major Depressive Disorder, OCD, and bulimia controlled clinical trials and US plus non-US Panic Disorder controlled clinical trials.” trials: “Multiple doses of PROZAC have been administered to 10,782 patients with various diagnoses in US clinical trials. In addition, there have been 425 patients administered PROZAC in panic clinical trials. Incidence in Major Depressive Disorder, OCD, bulimia, and Panic Disorder placebo-controlled clinical trials (excluding data from extensions of trials).” DailyMed

every row above was transcribed from the current FDA label on 2026-09-08; the full tables, including the reactions only one label lists, are on side-effect rates, drug vs placebo. the sexual side-effect rows carry each label's own underreporting caveat on the SSRI sexual side effects page.

escitalopram (Lexapro)

SSRI antidepressant

FDA approval: 2002 · full approval journey →
Who takes it (US, 2024): 9,460,411 patients · 41,800,787 prescriptions (AHRQ MEPS)2014 → 2024: 19,639,202 → 41,800,787 prescriptions, up 113%Correction (29 September 2026): An earlier version of this section took its numbers from ClinCalc DrugStats, a third-party site that republishes this survey; we now compute them from AHRQ’s public-use files ourselves. It showed 9,451,175 people and 41,773,078 prescriptions for 2024; the file gives 9,460,411 and 41,800,787.
Who pays (2024): Medicare: 12,700,559 claims · 2,728,707 beneficiariesMedicaid: ≥6,618,277 prescriptionsCommercial claims: not published in any public dataset — a real absence, not a zero.
Boxed warning: Yes — on the current FDA label. read it in context →
Adverse-event reports (FAERS): Caveat first: FAERS counts unverified REPORTS, not patients or rates — no denominator, and attention inflates reporting. Never compare these counts head-to-head; a longer-marketed, more-prescribed drug accumulates more reports by exposure alone. how to read FAERS →139,512 all-time reports; most-reported terms: fatigue, nausea, drug ineffective
Legal & safety record: the full litigation & recall record →

fluoxetine (Prozac)

SSRI antidepressant

FDA approval: 1987 · full approval journey →
Who takes it (US, 2024): 6,177,252 patients · 29,517,276 prescriptions (AHRQ MEPS)2014 → 2024: 23,867,189 → 29,517,276 prescriptions; the change is within the survey’s margin of errorCorrection (29 September 2026): An earlier version of this section took its numbers from ClinCalc DrugStats, a third-party site that republishes this survey; we now compute them from AHRQ’s public-use files ourselves. It said prescriptions were up 24% since 2014; that change is within the survey’s margin of error, so we no longer state a direction.
Who pays (2024): Medicare: 7,496,788 claims · 1,487,371 beneficiariesMedicaid: ≥6,359,320 prescriptionsCommercial claims: not published in any public dataset — a real absence, not a zero.
Boxed warning: Yes — on the current FDA label. read it in context →
Adverse-event reports (FAERS): Caveat first: FAERS counts unverified REPORTS, not patients or rates — no denominator, and attention inflates reporting. Never compare these counts head-to-head; a longer-marketed, more-prescribed drug accumulates more reports by exposure alone. how to read FAERS →131,850 all-time reports; most-reported terms: drug ineffective, nausea, fatigue
Legal & safety record: the full litigation & recall record →

questions for your prescriber, built from this page

this page does not tell you to start, stop or switch anything. it arms the conversation where that decision belongs. take these in:

which of Lexapro and Prozac did you have in mind first, and what in my history points that way?
the trials put these two close together on average. what, for me specifically, breaks the tie: another medication i take, a side effect i can't live with, sleep, weight, sexual function?
the label lists these reactions above placebo for the one you're suggesting. which of them should i expect in the first weeks, and which would make you want to hear from me the same day?
how long were the approval trials, and what do we know past that point?
how will we judge whether it is working, by when, and what is the plan if it isn't?
if i ever want to come off it, how is that done, and over how long?
is there a cheaper generic or a formulation that changes the dosing, and does my insurance prefer one of these two?

the longer checklist, for any antidepressant, is questions to ask before starting antidepressants. if you are already on one and it feels like it stopped working, that has its own page.

What the warnings mean, drug by drug

Each explainer reads one drug's current FDA label and the studies in our library, in both directions: the boxed warnings with the label's own numbers, the trials behind the approval and how long they ran, side effects by rate, and the research since.

Questions people search

Is Lexapro or Prozac better?

No honest page can answer that in one line. Head-to-head differences between comparable psychiatric medications are usually smaller than rankings suggest, and which fits you depends on your history, other medications, and side-effect trade-offs, a prescriber conversation. What can be compared is the public record (usage, payers, label warnings, approval history, litigation), which is what this page does, with sources.

Is Lexapro the same as escitalopram, and Prozac the same as fluoxetine?

Lexapro is a brand name for escitalopram. Prozac is a brand name for fluoxetine. escitalopram is a SSRI antidepressant; fluoxetine is a SSRI antidepressant. This page uses the generic names because the federal datasets do.

Which has fewer side effects, Lexapro or Prozac?

The two FDA labels cannot be compared head to head, in their own words, because each drug's trials ran in different people under different rules. What each label prints: Lexapro (escitalopram): 6% stopped for an adverse reaction vs 2% on placebo (Lexapro label, Revised:4/2024 (as printed in the SPL)). Prozac (fluoxetine): the label prints no discontinuation rate in the section we transcribed. The page shows the reactions both labels list, each against its own placebo column. In the 2018 Lancet network meta-analysis, all-cause dropout against placebo was 0.90 for escitalopram and 0.88 for fluoxetine.

Is Lexapro or Prozac more effective for depression?

In the 2018 Lancet network meta-analysis of 522 trials and 116,477 adults with major depressive disorder, both beat placebo on response at about 8 weeks: escitalopram odds ratio 1.68 (95% CrI 1.50 to 1.87), fluoxetine 1.52 (1.40 to 1.66). The authors described the differences between drugs as mostly modest, with wide intervals. The 2009 Cochrane review pooled the three trials with response data and found no difference in response at 6 to 12 weeks (odds ratio 0.81, 95% CI 0.60 to 1.10) or remission (0.86, 0.65 to 1.15), with a small edge for escitalopram on the depression-score change; the labels differ mostly in kinetics: fluoxetine’s active metabolite lasts 4 to 16 days, so its label requires 5 weeks before an MAOI against 14 days for escitalopram, and fluoxetine’s warnings section carries a QT prolongation item that escitalopram’s does not. The certainty of evidence was moderate to very low, 78% of the trials were industry funded, and the pooled effect of all antidepressants against placebo was a standardised mean difference of 0.30. None of that says which one fits you.

Which is prescribed more, Lexapro or Prozac?

The page shows federal survey estimates we compute from AHRQ’s Medical Expenditure Panel Survey public-use files for both escitalopram and fluoxetine: patients and prescriptions in the latest reliable year, and the yearly series from 2014 where AHRQ’s reliability rule allows it, with a direction stated only when the change is larger than the survey’s margin of error. Prescription volume reflects habits, prices, and marketing as much as evidence, so it is context, not a verdict.

Where these numbers come from

Patients and prescriptions: computed by us from AHRQ's public-use Medical Expenditure Panel Survey Prescribed Medicines files, 2014 (HC-168A) to 2024 (HC-254A), with AHRQ's reliability rule (a year with fewer than 60 respondents or a relative standard error over 50% is not shown) — counts of people filling prescriptions, not outcomes. A direction is stated only when the change is larger than the survey's margin of error. Until 29 September 2026 these came from ClinCalc DrugStats, a third-party site; see the correction under each drug where a figure changed.
Medicare/Medicaid: CMS public datasets; commercial insurance publishes no per-drug series, so it renders as an absence, never a synthesized number.
Label facts: the drugs' current FDA labels via DailyMed — quoted in context on each drug's approval-journey page.
FAERS: openFDA — reports, never rates; the caveat travels above every count.
Litigation: court records and regulator documents only, on the linked record pages; settled is never adjudicated.

Explore both drugs in depth: /m/escitalopram · /m/fluoxetine · or the whole library →