Prescribing Citalopram

Why Citalopram Still Matters

Citalopram is the SSRI that ran the largest antidepressant trial ever conducted. When the NIMH designed STAR*D and needed one drug to put 4,041 real-world depressed patients on as the mandatory first step, they picked citalopram, and they picked it for reasons that still make it a sound first-line choice today: minimal drug interactions and one of the mildest discontinuation profiles of any SSRI. It is cheap, generic, once-daily, and unremarkable in the best sense: it does its job without inducing or inhibiting the CYP450 enzymes that make fluoxetine and fluvoxamine a headache in anyone on polypharmacy.

Then, in 2011, the FDA issued a warning about dose-dependent QTc prolongation, capped the dose at 40 mg (20 mg in the elderly and in poor metabolizers), and citalopram's reputation took a hit it has never fully recovered from. Escitalopram, its own S-enantiomer, marketed as "the son of Celexa," inherited the franchise, with half the QTc effect and a marketing budget citalopram lacked.

Here is the thing worth knowing: the QTc warning is real pharmacology, but its clinical consequences have proven far smaller than the warning implied. Multiple large retrospective cohorts (Tennessee Medicaid, VA) found no increase in sudden cardiac death with high-dose citalopram versus other SSRIs, and the VA's forced deprescribing after the warning was associated with more hospitalizations and deaths, not fewer arrhythmias.

Part 1: Indications: Who Is Citalopram For?

FDA-Approved Use

• Major depressive disorder (adults)

That is the only FDA indication. Everything else is off-label but well-supported, because SSRIs are a class of broadly effective anxiolytics and antidepressants and citalopram behaves like its peers.

Evidence-Based Off-Label Uses

• Generalized anxiety disorder, panic disorder, social anxiety disorder: SSRIs are first-line; citalopram works as well as its class-mates.
• OCD: SSRIs match clomipramine for efficacy with far better tolerability; OCD typically needs the higher end of dosing and a longer trial (10–12 weeks).
• PTSD: reasonable off-label option; sertraline and paroxetine carry the formal indications, but the class effect applies.
• Agitation in Alzheimer's dementia (BPSD): a genuinely useful, evidence-backed niche. In the CitAD trial (Porsteinsson, JAMA 2014; 186 non-depressed Alzheimer's patients), citalopram 10 mg titrated to a max of 30 mg significantly reduced agitation versus placebo. Note the lower ceiling: these are elderly patients, and QTc and cognitive worsening at 30 mg were real signals in that study.

Who Citalopram Is Especially Good For

• The polypharmacy patient. Because citalopram neither induces nor inhibits CYP450 enzymes to any clinically meaningful degree, it is one of the safest antidepressants to layer onto a complex medication list: cardiac drugs, antiepileptics, transplant meds, oncology regimens. This is its signature advantage.
• The patient you may need to stop later. Its mild discontinuation profile makes it easier to taper than paroxetine or venlafaxine.
• The older adult (with the caveats below): alongside sertraline and escitalopram, citalopram is a preferred SSRI in geriatric depression for its clean interaction profile and lack of anticholinergic burden.
• The patient who needs bupropion augmentation. Citalopram plus bupropion is a clean, well-tolerated pairing (no CYP tangle), and it has among the best evidence of the SSRI-plus-bupropion combinations.

Who Citalopram Is Not For

• Bipolar depression as monotherapy. STAR*D excluded bipolar disorder for a reason: unopposed SSRIs can precipitate mood switching or accelerate rapid cycling. If you must use an antidepressant in bipolar depression, do it under cover of a mood stabilizer.
• Anyone with baseline QTc >500 ms, or on multiple other QT-prolonging drugs: choose a different SSRI (sertraline is the usual pivot).
• Complicated grief. Citalopram alone did not beat placebo; grief-focused psychotherapy is the effective treatment. (The trial was also hobbled by a mid-study forced dose reduction, but the practical message stands: don't reach for citalopram as a grief drug.)

Part 2: Before You Start

Citalopram requires no routine baseline labs in a healthy adult. The workup is short and mostly about identifying the small group in whom QTc matters.

The Pre-Start Checklist

QT-risk factors to tally: older age, female sex, structural heart disease, bradycardia, hypokalemia, hypomagnesemia, congenital long-QT, and concurrent QT-prolonging drugs (many antipsychotics, methadone, high-dose ondansetron, certain antibiotics/antiarrhythmics). A young, healthy adult with none of these needs neither an ECG nor much worry about the ceiling.

Part 3: How to Start and Dose

Formulation

Generic citalopram tablets (10, 20, 40 mg) and an oral solution (10 mg/5 mL) for fine titration. Citalopram is a racemic mixture: a 50/50 blend of R- and S-enantiomers, of which only the S-enantiomer (escitalopram) is pharmacologically active at the serotonin transporter (the S-form is far more potent at SERT). This is why 20 mg of citalopram is roughly equivalent to 10 mg of escitalopram.

Starting Dose and Titration

• Start: 20 mg once daily, morning or evening (take at night if it's sedating, in the morning if it's activating or disrupts sleep). In the anxious or elderly patient, 10 mg for the first week softens the early jitteriness and nausea, then move to 20 mg.
• The single most important dosing move: get to 40 mg. Titrate to 40 mg/day at ~week 4 if response is partial or absent and the patient is tolerating it. STAR*D's effective average was 41.8 mg: 20 mg is frequently a sub-therapeutic parking spot.
• 60 mg is off the table for routine use since the 2011 FDA action (originally "should not be used above 40 mg," softened in 2012 to "not recommended"). The efficacy gain above 40 mg is small and the QTc cost climbs; in the rare treatment-resistant patient where you consider it, do so deliberately, with an ECG, and document your reasoning.
• Elderly / hepatic impairment / poor CYP2C19 metabolizers / strong CYP2C19 inhibitors: cap at 20 mg/day.

The Therapeutic Ceiling Table: Memorize This

Onset and the Trial-Length Discipline

• Serotonin transporter blockade is immediate, but the antidepressant effect takes ~4 weeks to begin and often 6–7 weeks to mature (the delay is thought to reflect 5-HT1A autoreceptor desensitization and downstream BDNF/synaptic remodeling, not simply the acute rise in synaptic serotonin).
• Steady state: ~1 week (half-life ~35 hours, long enough for comfortable once-daily dosing and forgiving of a missed dose).
• Do not judge the trial before 4 weeks at 40 mg, and give a full trial 6–8 weeks before declaring failure. The most common prescribing error is bailing early.

Part 4: Monitoring

Citalopram needs little routine monitoring in a healthy adult: no drug levels, no mandatory labs. Focus your attention where it actually pays off.

The Monitoring Schedule

Targeted Checks (Not for Everyone)

• ECG: only in QT-risk patients, at baseline, after reaching 40 mg, or when adding another QT drug. Discontinue (or don't exceed) citalopram if QTc >500 ms. For a healthy young adult, routine ECG is unnecessary.
• Sodium: check in older adults if any confusion, lethargy, dizziness, or new falls: SSRI hyponatremia most often appears within the first ~2 weeks.
• Bleeding: no lab, but stay alert in patients on NSAIDs/anticoagulants; consider a PPI, which effectively neutralizes the SSRI GI-bleed risk.

Part 5: Side Effects and How to Manage Them

The governing principle is the SSRI principle: early side effects are usually transient and warrant reassurance; the durable side effects are sexual dysfunction and emotional blunting, both of which are dose-related and manageable. Citalopram is, on balance, one of the better-tolerated SSRIs.

Early / Transient (First 1–2 Weeks)

Nausea, loose stools, headache, jitteriness, transient insomnia or sedation. These typically resolve within 3–7 days.

• Take with food for nausea.
• Move the dose (evening if sedating, morning if activating).
• Reassure: most patients who quit in week 1 quit over side effects that would have faded.

Gastrointestinal

Citalopram is more prone to diarrhea than average (odds ratio ~1.64 across antidepressants) but is less likely to cause constipation than SNRIs, a tolerability plus.

• Persistent diarrhea: dose with food, consider PRN loperamide, or switch agents if intolerable.

Sexual Dysfunction: The Durable One

The most common lasting SSRI side effect (~50–70% across the class): decreased libido, delayed or absent orgasm, erectile difficulty, genital numbing. Often under-reported unless you ask directly.

• Ask at every follow-up: patients rarely volunteer it and it drives silent non-adherence.
• Lower the dose if efficacy allows (often helps without losing benefit).
• Switch agent: not every SSRI produces the same profile in the same person; if two SSRIs both cause it, consider a non-SSRI.
• Bupropion: either as a switch or as add-on, has the least sexual dysfunction and can offset the SSRI effect.
• Sildenafil/tadalafil for the erectile component.

Emotional Blunting

A gray, flattened, "zombie" feeling with loss of positive as well as negative emotion. This is a signal of too-high a dose, not an expected therapeutic state.

• Lower the dose. If the patient can't find a dose that treats the depression without blunting, that's a reason to change drugs.

Weight

SSRI-associated weight gain runs ~25% as a class, often modest and partly reflecting recovered appetite. Citalopram is metabolically neutral relative to mirtazapine and TCAs.

• Diet/exercise counseling; consider a more weight-neutral agent (or bupropion) if it becomes a dealbreaker.

Sleep / Energy

Usually mild. If activating, dose in the morning; if sedating, dose at night. Persistent fatigue: rule out relapse and hypothyroidism before blaming the drug.

Other

• Sweating, mild tremor: usually minor and dose-related.
• Hyponatremia (see Part 6): mostly older adults.
• Bruising / bleeding tendency: via platelet serotonin depletion.
• Sedation and cognitive fog are less a citalopram problem than a paroxetine problem; if a patient is foggy on citalopram, check the thyroid and the dose before assuming the drug.

Part 6: Overdose, Toxicity, and the QTc Question

Overdose

SSRIs are, as a class, relatively safe in overdose compared with TCAs and MAOIs, a meaningful advantage in a depressed, potentially suicidal population. The citalopram-specific caveat is QTc prolongation and the risk of torsades in large overdoses, which is more pronounced than with other SSRIs and can be delayed. Any significant citalopram overdose warrants ED evaluation with cardiac monitoring and serial ECGs. As with any SSRI, watch for serotonin syndrome, especially with co-ingestants.

Serotonin Syndrome

Agitation, confusion, autonomic instability (tachycardia, hyperthermia, diaphoresis), hyperreflexia, clonus, rigidity. Risk rises when citalopram is combined with MAOIs (absolutely contraindicated), and to a much lesser degree with tramadol, linezolid, methylene blue, triptans, or other serotonergics. Management: stop the offending agents, supportive care, cyproheptadine in severe cases. In practice, the triptan risk is very low, on the order of 0.6–2.3 cases per 10,000 person-years, and a migraineur on an SSRI who occasionally uses a triptan does not need to be denied either.

The QTc Warning: What It Means and What It Doesn't

This is citalopram's defining safety issue, so be precise about it:

• The effect is real and dose-dependent. Citalopram 60 mg prolongs QTc by roughly 18.5 ms; escitalopram 30 mg by about half that (~10.7 ms). Citalopram is the most QT-prolonging SSRI.
• The FDA response (2011, softened 2012): don't exceed 40 mg generally, 20 mg in the elderly/hepatic/poor-metabolizer/strong-CYP2C19-inhibitor groups; stop if QTc >500 ms.
• But the clinical outcomes have not matched the warning's alarm. Large retrospective cohorts, Ray (Tennessee Medicaid, 54,220 high-dose SSRI users) and Zivin (VA), found no significant increase in sudden cardiac death or ventricular arrhythmia with high-dose citalopram versus other SSRIs. And when the VA forced roughly 35,000 veterans to lower their dose after the warning, the result was a rise in all-cause hospitalizations and deaths with no decline in arrhythmias.

Part 7: Drug Interactions

Citalopram's low interaction burden is one of its best features, but it is not zero, and the interactions it does have cluster around two axes: CYP2C19 (which changes its own levels) and QTc (which changes its safety).

The Interactions That Raise Citalopram Levels (Cap at 20 mg)

Citalopram is metabolized primarily by CYP2C19. Strong 2C19 inhibitors raise its levels and, with them, the QTc risk:

• Omeprazole / esomeprazole (PPIs)
• Cimetidine (H2 blocker)
• Fluconazole, fluvoxamine, and other strong 2C19 inhibitors

When these are co-prescribed (or the patient is a genetic poor 2C19 metabolizer), hold the ceiling at 20 mg. A patient on chronic omeprazole is functionally in the same bin as an elderly patient.

The Interaction That Matters Most Clinically: Additive QTc

Combining citalopram with other QT-prolonging drugs (antipsychotics such as ziprasidone, haloperidol, thioridazine, methadone, class IA/III antiarrhythmics, certain macrolide/fluoroquinolone antibiotics, high-dose IV ondansetron) stacks the QTc effect. This is the interaction to actively screen for. Check an ECG if you must combine, correct hypokalemia/hypomagnesemia, and consider a non-citalopram SSRI if the QT burden is high.

Bleeding

SSRIs plus NSAIDs raise GI-bleed risk substantially (on the order of 7–15 fold in higher-risk analyses); SSRIs plus warfarin/antiplatelets also elevate it. Co-prescribe a PPI in high-risk patients; it drops the excess bleed risk to negligible. Note the elegant catch: omeprazole both protects against the bleed and raises citalopram levels, so if you use it for GI protection, keep the citalopram at 20 mg.

MAOIs: Absolute Contraindication

Never co-administer. Washout 14 days in either direction (from an MAOI to citalopram, and from citalopram to an MAOI, given its ~35-hour half-life).

What Citalopram Does Not Meaningfully Do

• It does not inhibit or induce CYP2D6, 3A4, or 1A2 to a clinically important degree. This is the whole point of the drug. Unlike fluoxetine (potent 2D6/2C19 inhibitor, raises TCAs, aripiprazole, risperidone, and can block activation of prodrugs like tamoxifen, tramadol, and codeine; effects linger for weeks) or fluvoxamine (potent 1A2/3A4 inhibitor, dangerous with clozapine, theophylline, tizanidine), citalopram leaves other drugs' levels essentially untouched.

Part 8: Special Populations

Pregnancy

The mainstream position: for moderate-to-severe depression, treating usually beats not treating, because untreated antenatal depression carries real fetal and maternal risk, partly through behavioral pathways (missed prenatal care, poor nutrition, substance use) and partly through relapse and postpartum depression.

• First-trimester teratogenicity is low; the class signal for cardiac septal defects is small and most concentrated with paroxetine, which is generally avoided in pregnancy.
• Third trimester: a self-limited neonatal adaptation syndrome (jitteriness, tremor, feeding/respiratory difficulty) occurs in a minority of exposed newborns, and there is a small absolute increase in persistent pulmonary hypertension of the newborn. Neither justifies reflexive discontinuation of an effective medication in a woman who needs it.
• Escitalopram is sometimes preferred over citalopram in pregnancy (S-enantiomer only, lower total exposure), but citalopram is a reasonable, well-studied choice, especially if a woman is already stable on it.
• Practical rule: individualize with the patient and OB, don't stop a working antidepressant out of reflex, and if you continue, plan neonatal observation.

Lactation

Citalopram is excreted into breast milk in low amounts and is generally considered compatible with breastfeeding; sertraline and paroxetine have the lowest milk levels if you're choosing de novo, but an infant of a stably treated mother rarely needs more than routine pediatric observation.

Elderly (≥60)

• Cap at 20 mg/day (QTc safety).
• Start low (often 10 mg) and go slow.
• Hyponatremia is the population-specific hazard, highest in older adults, usually within the first 2 weeks; check sodium if confused, lethargic, or falling.
• Also watch bleeding risk (screen NSAIDs/anticoagulants), falls/fractures (SSRIs modestly reduce bone density), and stack QT drugs cautiously.
• The upside: citalopram's clean interaction profile and lack of anticholinergic load make it a preferred SSRI in geriatric depression.

Renal and Hepatic Impairment

• Renal: most SSRIs, citalopram included, are safe even in significant renal impairment (hepatic clearance dominates); no routine adjustment, monitor in ESRD.
• Hepatic: cap at 20 mg; reduced clearance raises levels and QTc risk.

Children and Adolescents

• Fluoxetine is the preferred first-line pediatric SSRI (best studied, FDA-approved for depression down to age 8). Citalopram has some supporting evidence but is second-line.
• The boxed suicidality warning (ages <25) is real to monitor but grounded in a modest ~2% absolute difference (4% vs 2%) with no completed suicides in the pediatric trials; a leading interpretation is increased disclosure rather than induced ideation. Monitor closely in the first 2–4 weeks; don't let the warning scare you away from treating a depressed adolescent, given that the 2004 warning was followed by a drop in prescribing and a rise in adolescent suicide.

Part 9: Discontinuation

Citalopram has one of the mildest discontinuation profiles among SSRIs: a real advantage, and part of why it was chosen for STAR*D. It is nowhere near paroxetine or venlafaxine on the withdrawal spectrum (its ~35-hour half-life gives a gentler taper than paroxetine's shorter one, though not the built-in self-taper of fluoxetine).

• Discontinuation syndrome (when it occurs): dizziness, nausea, fatigue, flu-like malaise, irritability, and the characteristic "brain zaps" (brief, non-painful electric-shock sensations). Roughly 60% of patients have no withdrawal, ~35% mild-to-moderate, ~5% severe, with citalopram skewing toward the benign end.
• Risk factors for a rougher stop: high dose, long duration (years), abrupt cessation.
• How to taper: for short courses (a few months), a modest taper over 2–4 weeks usually suffices. For long-term or higher-dose users, taper more gradually, hyperbolically, taking smaller and smaller decrements as you approach zero (e.g., 40 → 20 → 10 → 5 → stop, using the oral solution for fine steps if needed). The iterative rule: if withdrawal symptoms emerge, go back to the last comfortable dose and taper more slowly. With patience this essentially never fails.
• Rapid discontinuation needed? Citalopram tolerates it better than most, but if you truly need a clean abrupt stop, fluoxetine's long half-life makes it the safer bridge.

Part 10: Citalopram vs the Alternatives

• vs Escitalopram (Lexapro): escitalopram is citalopram's active S-enantiomer. It offers a marginal (~5–10%) efficacy edge in equipotent head-to-head trials, possibly slightly faster onset, and, most importantly, about half the QTc prolongation. Two population studies found more cardiac arrests/serious arrhythmias with citalopram than escitalopram. If cost is no object and QT is any concern, escitalopram is the cleaner choice. Citalopram remains fully valid, especially where the generic price or formulary favors it. To switch: add escitalopram at 5 mg and titrate to roughly half the citalopram dose over 2–4 weeks while tapering the citalopram.
• vs Sertraline: comparable efficacy; sertraline has no QTc ceiling and is a common pivot when QT is the concern, at the cost of somewhat more diarrhea. Both are geriatric- and pregnancy-friendly.
• vs Fluoxetine: fluoxetine is more activating (good for the fatigued patient, bad for the insomniac), has a long half-life (built-in taper, but a 5-week MAOI washout), and is a potent CYP inhibitor, the opposite of citalopram's clean interaction profile. Fluoxetine is preferred in pediatrics.
• vs Paroxetine: avoid paroxetine where you can: worst sexual dysfunction, worst withdrawal, anticholinergic, cognitive fog, pregnancy cardiac signal. Citalopram beats it on nearly every tolerability axis.
• vs Fluvoxamine: fluvoxamine's potent 1A2/3A4 inhibition makes it interaction-laden; reserved mostly for OCD. Citalopram is far friendlier in polypharmacy.
• vs SNRIs (venlafaxine, duloxetine): SNRIs may edge citalopram at high doses in some analyses but add blood-pressure elevation (venlafaxine) and worse discontinuation. STAR*D found no efficacy advantage switching citalopram non-remitters to venlafaxine.
• vs Bupropion: roughly equal antidepressant efficacy, even with comorbid anxiety. Bupropion wins on sexual side effects (it's sexual-sparing) and pairs cleanly with citalopram as an augment, the best-evidenced SSRI-plus-bupropion combination.
• vs TCAs / MAOIs: citalopram is far better tolerated and vastly safer in overdose; TCAs and MAOIs are later-line for genuine treatment resistance.

Mechanism: The Short Version

Citalopram is a highly selective serotonin reuptake inhibitor: it blocks the serotonin transporter (SERT), leaving more serotonin in the synapse, with negligible noradrenergic, dopaminergic, histaminergic, or anticholinergic activity (which is why it's clean on sedation, weight, and cognition relative to older agents). Transporter blockade is immediate, but the antidepressant effect emerges over ~4 weeks, attributed to gradual desensitization of 5-HT1A autoreceptors and downstream BDNF-driven synaptic remodeling rather than to the acute rise in synaptic serotonin. It is a racemate; only the S-enantiomer (escitalopram) is active at SERT, with the R-enantiomer largely inert and possibly mildly antagonistic to the S-form's effect, the pharmacologic rationale behind marketing the purified S-enantiomer separately. Half-life is ~35 hours; metabolism is primarily hepatic via CYP2C19 (with minor 3A4/2D6 contributions), and it neither meaningfully induces nor inhibits CYP enzymes.

The Bedside Cheat Sheet