desvenlafaxine (Pristiq)

the boxed warning: suicidal thoughts and behaviors under 25

the box reads: "antidepressants increased the risk of suicidal thoughts and behavior in children, adolescents, and young adults in short-term studies. these studies did not show an increase in the risk of suicidal thoughts and behavior with antidepressant use in patients over age 24; there was a reduction in risk with antidepressant use in patients aged 65 and older." it tells prescribers to monitor everyone starting an antidepressant closely for worsening and for new suicidal thoughts, to advise families and caregivers to watch and to talk to the prescriber, and it adds that desvenlafaxine "is not approved for use in pediatric patients" [label, boxed warning].

the numbers behind it are the fda’s pooled analyses of the whole class, not desvenlafaxine alone: 24 short-term trials of 9 antidepressants in more than 4,400 children and adolescents, and 295 short-term trials of 11 antidepressants in more than 77,000 adults. per 1,000 people treated, the drug-placebo difference was 14 additional cases of suicidality under 18, 5 additional cases from 18 to 24, 1 fewer case from 25 to 64, and 6 fewer at 65 and older. there were no suicides in the pediatric trials, and too few in the adult trials to reach any conclusion about suicide itself [label 5.1]. in the youth trials, suicidality was reported in about 4% on drug against 2% on placebo [boxed-warning-row]. the label says the closest watch belongs in "the initial few months" and "at times of dose changes, either increases or decreases", and it names the signs that may come first: anxiety, agitation, panic attacks, insomnia, irritability, hostility, impulsivity, akathisia, hypomania and mania [label 5.1].

what the label says about its own trials, arm by arm

the approval rests on four 8-week placebo-controlled trials in adult outpatients, scored on the 17-item hamilton depression scale, where people started at about 23 to 25 points [label 14, table 9]. the differences from placebo at the end, in points:

• study 1 (100, 200 or 400 mg, about 115 people per arm, 118 on placebo): 100 mg was 2.9 points better than placebo (95% confidence interval 0.8 to 5.1), 200 mg 2.0 points, which did not reach statistical significance, and 400 mg 3.1 points (0.9 to 5.2) [label 14, table 9].
• study 2 (200 or 400 mg, 121 and 124 people, 124 on placebo): 3.3 points (1.2 to 5.3) and 2.8 points (0.7 to 4.8) [label 14, table 9].
• study 3 (50 or 100 mg, 150 and 147 people, 150 on placebo): 50 mg 1.9 points (0.3 to 3.5); 100 mg 1.5 points, which did not reach statistical significance [label 14, table 9].
• study 4 (50 or 100 mg, 164 and 158 people, 161 on placebo): 2.5 points (0.9 to 4.1) and 3.0 points (1.4 to 4.7) [label 14, table 9].
• a postmarketing study (25 or 50 mg, 232 and 236 people, 231 on placebo): 50 mg beat placebo; 25 mg did not [label 14].

what those trial numbers mean, and do not

the label says desvenlafaxine "showed superiority over placebo" on the hamilton scale "in four studies" and on the clinicians’ overall improvement rating in three of the four [label 14]. both are true of each study taken whole; table 9 also shows that two of the nine dose arms in those four studies did not separate from placebo on their own [label 14, table 9]. every significant difference sits between 1.9 and 3.3 points on a 52-point scale, and the lower ends of the confidence intervals run as low as 0.3 points [label 14, table 9]. the label does not describe trials that failed outright, so the four shown are the ones that supported approval; the only negative result it prints is the 25 mg arm of the later study [label 14].

on dose, the label is direct: "in studies directly comparing 50 mg per day and 100 mg per day there was no suggestion of a greater effect with the higher dose and adverse reactions and discontinuations were more frequent at higher doses" [label 14]. the dropout figures show what that means: 4.1% stopped for a side effect at 50 mg against 3.8% on placebo, 8.7% at 100 mg, and 12% across all doses from 50 to 400 mg against 3% on placebo [label 6.1]. the 25 mg tablet exists for tapering and for severe kidney disease, not as a treatment dose [label 2.1, 2.2].

on staying on it, two relapse trials took people who had responded and either kept them on desvenlafaxine or switched them to placebo for up to 26 weeks. at 50 mg, an estimated 14% relapsed on the drug against 30% on placebo; at 200 or 400 mg, 29% against 49% [label 14]. that design shows what happens when a drug that is working is withdrawn, which is a different question from how well it works in the first place; switching to placebo is also a way of stopping the drug, which the label warns has its own symptoms [label 14; label 5.7].

in children, two 8-week placebo-controlled trials in 587 people aged 7 to 17 did not show efficacy, and weight loss of 3.5% or more of body weight occurred in 22% on the low dose, 14% on the high dose and 7% on placebo [label 8.4].

the other warnings, in plain words

beyond the boxed warning, section 5 lists:

• serotonin syndrome: "a potentially life-threatening condition", more likely with other serotonergic drugs (the label names triptans, tricyclics, fentanyl, lithium, tramadol, meperidine, methadone, tryptophan, buspirone, amphetamines and st john’s wort) and with maois, but it "can also occur when these drugs are used alone" [label 5.2]. maois are contraindicated within 14 days before and 7 days after desvenlafaxine, and it must not be started in anyone being given linezolid or intravenous methylene blue [label 4, 2.7, 2.8]. the signs a clinician looks for are inducible muscle jerking, agitation, sweating, tremor and overactive reflexes, and in severe cases rigidity and fever [hunter-2003].
• blood pressure: "increases in blood pressure were observed in clinical studies", high blood pressure should be controlled before starting, and blood pressure should be checked regularly [label 5.3]. in the short-term trials, sustained high diastolic pressure (90 or more, and at least 10 above baseline, at three visits in a row) occurred in 1.3% at 50 mg, 0.7% at 100 mg, 1.1% at 200 mg and 2.3% at 400 mg, against 0.5% on placebo; average lying systolic pressure rose 1.2 to 2.5 points on the drug and fell 1.4 on placebo [label 6.1, tables 6 and 7].
• bleeding: drugs that block serotonin reuptake "may increase the risk of bleeding events", more so with aspirin, nsaids, warfarin and other anticoagulants, and exposure in the month before delivery has been associated with a less than twofold increase in postpartum hemorrhage [label 5.4].
• angle-closure glaucoma: the pupil dilation antidepressants cause can trigger an attack in someone with untreated narrow angles [label 5.5]. in the trials, dilated pupils occurred in 2% at 50 and 100 mg and 6% at the higher doses, against under 1% on placebo [label 6.1, table 2].
• mania: reported in about 0.02% of people in the trials; caution with a personal or family history of mania or hypomania, and screening for bipolar disorder before starting, because the drug "is not approved for use in treating bipolar depression" [label 5.6, 5.1].
• seizures: reported in the trials, which excluded people with a seizure history, so it "has not been systematically evaluated" in seizure disorders [label 5.8].
• low sodium: hyponatremia, often from inappropriate antidiuretic hormone, with cases below 110 mmol/l; older people and people on diuretics are at greater risk, and the early signs are headache, poor concentration, memory problems, confusion, weakness and unsteadiness [label 5.9].
• lung: interstitial lung disease and eosinophilic pneumonia have rarely been reported with venlafaxine, the parent drug; progressive breathlessness, cough or chest discomfort calls for prompt evaluation [label 5.10].
• sexual function: snris "may cause symptoms of sexual dysfunction", and the label asks prescribers to ask about it before and during treatment "because sexual function may not be spontaneously reported" [label 5.11]. the numbers are in the side-effect section below.
• older people: systolic orthostatic hypotension, a drop of 30 points or more on standing, occurred in 8% (7 of 87) of people 65 and over on desvenlafaxine against 2.5% (1 of 40) on placebo, and in under 1% of younger people on either [label 6.1, 8.5]. only 6% of the 4,158 people in the pre-approval trials were 65 or older [label 8.5].
• kidneys and liver: the maximum is 50 mg a day with moderate kidney impairment and 25 mg a day (or 50 mg every other day) with severe impairment or dialysis; with moderate to severe liver impairment, doses above 100 mg are not recommended [label 2.2, 2.3].
• pregnancy and breastfeeding: there are no published studies of desvenlafaxine in pregnant women; studies of venlafaxine have not shown a clear link to birth defects, but claims studies link it to preeclampsia (adjusted relative risk 1.57, 1.29 to 1.91, with exposure in mid pregnancy) and to postpartum hemorrhage with exposure through delivery (2.24, 1.69 to 2.97), both open to confounding by the depression itself. newborns exposed late in pregnancy can need breathing support and tube feeding. in a study of 10 breastfeeding women, the infant received about 6.8% of the mother’s weight-adjusted dose, with no adverse reactions seen [label 8.1, 8.2]. the label also notes that in a study of 201 women, those who stopped antidepressants in pregnancy were more likely to relapse [label 8.1].
• interactions: desvenlafaxine raises levels of drugs cleared by cyp2d6 (the label names desipramine, atomoxetine, dextromethorphan, metoprolol, nebivolol, perphenazine and tolterodine), enough at 400 mg that their doses may need halving, but no change is needed at 100 mg or below [label 7]. its own clearance does not depend on cyp2d6 [label 12.3]. it is not a controlled substance [label 9.1], and it can cause false-positive urine immunoassays for pcp and amphetamine [label 7].

what the label says about stopping

quoted so that it is the label’s and not ours: "gradually reduce the dosage rather than stopping desvenlafaxine extended-release tablets abruptly when discontinuing therapy". and: "if intolerable symptoms occur following a decrease in the dose or upon discontinuation of treatment, then resuming the previously prescribed dose may be considered. subsequently, the healthcare provider may continue decreasing the dose, but at a more gradual rate. in some patients, discontinuation may need to occur over a period of several months" [label 2.5, 5.7]. the listed symptoms are nausea, sweating, low mood, irritability, agitation, dizziness, electric-shock sensations, tremor, anxiety, confusion, headache, lethargy, emotional swings, insomnia, hypomania, tinnitus and seizures [label 5.7].

this label goes further than most. it says there have been postmarketing reports of serious discontinuation symptoms "which can be protracted and severe", and that "completed suicide, suicidal thoughts, and severe aggression (including hostility, rage, and homicidal ideation) have been observed in patients during reduction in desvenlafaxine extended-release tablets dosage, including during discontinuation", along with blurred vision and raised blood pressure after stopping or cutting the dose [label 5.7]. it also warns that switching from venlafaxine to desvenlafaxine has caused discontinuation symptoms, so the first drug may need tapering [label 2.6]. the uk guideline on dependence-forming medicines says withdrawal "may take several months or more" and that taper steps should shrink as the dose falls [nice-2022]; the royal college of psychiatrists said in 2019 that for some people symptoms "last much longer and are more severe" than guidelines had said [rcpsych-2019].

side effects, with the label’s numbers

from the pooled 8-week trials, rates at 50 mg and 100 mg against placebo, with the 400 mg rate in brackets to show the dose gradient: nausea 22% and 26% against 10% (41%); dizziness 13% and 10% against 5% (16%); dry mouth 11% and 17% against 9% (25%); insomnia 9% and 12% against 6% (15%); excessive sweating 10% and 11% against 4% (21%); constipation 9% and 9% against 4% (14%); fatigue 7% and 7% against 4% (11%); decreased appetite 5% and 8% against 2% (10%); somnolence 4% and 9% against 4% (12%); anxiety 3% and 5% against 2%; blurred vision 3% and 4% against 1% [label 6.1, table 2]. in the trials, total cholesterol rose past a clinically notable threshold in 3% at 50 mg and 10% at 400 mg against 2% on placebo, and trace protein in the urine appeared in 6% at 50 mg and 8% at 100 mg against 4%, without kidney changes and generally transient [label 6.1, tables 4 and 5]. ischemic heart events were uncommon but more frequent on the drug than placebo, in people with several cardiac risk factors [label 6.1].

sexual side effects, men, at 50 mg and 100 mg against placebo (400 mg in brackets): erectile dysfunction 3% and 6% against 1% (11%); delayed ejaculation 1% and 5% against under 1% (6%); decreased libido 4% and 5% against 1% (3%); anorgasmia 0% and 3% against 0% (8%). in women, anorgasmia was 1% at both doses against 0% (3%) [label 6.1, table 3]. these are spontaneous reports in 8-week trials; the label itself says sexual function "may not be spontaneously reported" [label 5.11].

two cautions the label attaches to every figure: rates from trials "cannot be directly compared to rates in the clinical studies of another drug and may not reflect the rates observed in clinical practice" [label 6.1].

what the research says, in both directions

in favour. in the largest comparison of antidepressants, 522 randomised trials in 116,477 adults, desvenlafaxine’s odds of response against placebo were 1.49 (credible interval 1.24 to 1.79), and its odds of dropping out for any reason were 1.08 (0.88 to 1.33), no different from placebo [cipriani-2018; the desvenlafaxine figures were read from the paper’s figure 3 for this site’s comparison pages]. the label’s own relapse trials halved the relapse rate at 50 mg over 26 weeks, 14% against 30% [label 14]. its pharmacology is simple: no dependence on cyp2d6, no qt prolongation in a thorough qt study, and a single dose that the label says is both the starting and the working dose [label 12.2, 12.3, 2.1].

against, or smaller than it sounds. in the same 522-trial comparison, desvenlafaxine is named in none of the abstract’s head-to-head groups, either the more effective or the less, and the authors graded the certainty of the evidence moderate to very low [cipriani-2018]. desvenlafaxine was one of four antidepressants approved between 2008 and 2013 whose fda files were compared with the journals: across 30 trials in 13,747 patients, every positive trial was published transparently but only 7 of 15 negative ones were, and the effect size in the fda’s own records was 0.24 (0.18 to 0.30) against 0.29 in the published papers, a small effect by convention [turner-2022]. the label’s significant results sit between 1.9 and 3.3 hamilton points, two of nine dose arms missed, a 25 mg dose failed, and two pediatric trials failed [label 14, table 9; label 8.4].

on sex after stopping, where the record is thinner than the label reads. in may 2019 the european regulator’s safety committee required desvenlafaxine, with nine other ssris and snris, to carry the warning that "there have been reports of long-lasting sexual dysfunction where the symptoms have continued despite discontinuation"; how often it happens is unknown [prac-2019]. the us label’s section 5.11 says nothing about persistence after stopping, either way [label 5.11].

questions people ask

what does the pristiq (desvenlafaxine) black box warning mean?

it says antidepressants increased the risk of suicidal thoughts and behavior in children, adolescents and young adults in short-term studies, did not increase it over age 24, and reduced it at 65 and older. per 1,000 people treated in the fda’s pooled trials, that was 14 additional cases under 18, 5 additional from 18 to 24, 1 fewer from 25 to 64 and 6 fewer at 65 and over. it asks for close monitoring of everyone starting an antidepressant, especially early and at dose changes. desvenlafaxine is not approved for anyone under 18. it is a warning about watching closely, not an instruction to stop.

is 100 mg of desvenlafaxine better than 50 mg?

not in the label’s trials. the label says that in studies directly comparing 50 and 100 mg there was "no suggestion of a greater effect with the higher dose", while side effects and discontinuations were more frequent at higher doses: 4.1% stopped for a side effect at 50 mg, 8.7% at 100 mg, against 3.8% on placebo. it calls 50 mg "both a starting dose and the therapeutic dose". in one of the two trials that tested both, 50 mg separated from placebo and 100 mg did not.

what is the difference between desvenlafaxine and venlafaxine?

desvenlafaxine is the main active metabolite of venlafaxine: the body converts venlafaxine into it. both are snris for depression. desvenlafaxine’s label sets one dose, 50 mg, does not depend on the cyp2d6 liver enzyme, and is contraindicated in anyone allergic to venlafaxine. its label reprints venlafaxine’s overdose record because its own is limited, and says switching from venlafaxine to desvenlafaxine has caused discontinuation symptoms. the side-by-side comparison on this site sets the two labels next to each other.

what happens if you stop desvenlafaxine suddenly?

the label lists nausea, sweating, low mood, irritability, agitation, dizziness, electric-shock sensations, tremor, anxiety, confusion, headache, insomnia, tinnitus and seizures, says postmarketing reports describe symptoms that "can be protracted and severe", and says completed suicide, suicidal thoughts and severe aggression have been observed during dose reduction and discontinuation. its instruction is a gradual reduction, stepping back to the previous dose if symptoms are intolerable, and it says stopping may take several months in some people. a plan belongs with a prescriber.

does desvenlafaxine raise blood pressure?

it can. the label says increases were seen in the trials and asks for regular monitoring and for high blood pressure to be controlled before starting. sustained high diastolic pressure occurred in 0.7% to 2.3% of people across doses against 0.5% on placebo, highest at 400 mg, and cases needing immediate treatment have been reported. in people 65 and over, the opposite problem, a large drop on standing, occurred in 8% against 2.5% on placebo.

what are the most common desvenlafaxine side effects?

at the 50 mg dose in the label’s 8-week trials: nausea 22% against 10% on placebo, dizziness 13% against 5%, dry mouth 11% against 9%, excessive sweating 10% against 4%, insomnia 9% against 6%, constipation 9% against 4%, and decreased appetite 5% against 2%. most rise with dose. in men, erectile dysfunction reached 11% at 400 mg and delayed ejaculation 7% at 200 mg, against 1% and under 1% on placebo.

sources for this section

the library rows link to our one-page summary of each study, with the original paper, its funding and our evidence grade on that page.

1. desvenlafaxine extended-release tablets prescribing information, fda label on dailymed (set id ff837eb1-24c3-4c63-9b4b-c6b5935a9b47, version 7, effective august 21, 2026; advagen, anda 204028): boxed warning; sections 1, 2.1 to 2.8, 4, 5.1 to 5.11, 6.1 (tables 2 to 7), 6.2, 7, 8.1, 8.2, 8.4, 8.5, 9.1, 10.1, 12.2, 12.3, 14 (table 9)
2. drugs@fda, nda 021992 (pristiq), approved february 29, 2008 as a new molecular entity
3. ahrq medical expenditure panel survey, prescribed medicines public-use files 2014 (hc-168a) through 2024 (hc-254a): desvenlafaxine was reported by 21 to 51 respondents a year, under ahrq’s 60, so no year is reportable; computed by veisund, method on the usage section of this page
4. the antidepressant class boxed warning on suicidal thoughts and behaviors, read for the library from the prozac label: the 24 pediatric trials, about 4% against 2%, and the age boundary (library row)
5. cipriani a, et al. comparative efficacy and acceptability of 21 antidepressant drugs for the acute treatment of adults with major depressive disorder. lancet 2018 (the library row is written up for venlafaxine; desvenlafaxine’s odds ratios were read from the paper’s figure 3 on 2026-10-11 for this site’s comparison pages) (library row)
6. turner eh, et al. selective publication of antidepressant trials and its influence on apparent efficacy: updated comparisons and meta-analyses of newer versus older trials. plos medicine 2022 (desvenlafaxine, vilazodone, levomilnacipran and vortioxetine) (library row)
7. european medicines agency, pharmacovigilance risk assessment committee. recommendations on signals adopted at the 13 to 16 may 2019 meeting, section 1.3: ssris and snris, persistent sexual dysfunction after drug withdrawal (desvenlafaxine named) (library row)
8. national institute for health and care excellence. medicines associated with dependence or withdrawal symptoms (ng215), 2022 (library row)
9. royal college of psychiatrists. position statement on antidepressants and depression (ps04/19), 2019 (library row)
10. dunkley ej, et al. the hunter serotonin toxicity criteria. qjm 2003 (library row)
11. what the venlafaxine (effexor) warnings mean, the companion explainer on this site for the parent drug
12. desvenlafaxine vs venlafaxine, the side-by-side comparison page on this site

Most-reported reactions

• Drug ineffective2,724
• Nausea2,439
• Headache1,943
• Dizziness1,895
• Anxiety1,818
• Fatigue1,677
• Insomnia1,574
• Feeling abnormal1,451
• Depression1,388
• Off label use1,054