Citalopram

Limits on what this shows

• Key negative/null findings: No human mitochondrial-function data for citalopram were located. The study used a single concentration (20 µmol/L) and a single disease-modeling cell line; it does not establish an effect in normal neurons or in non-neuronal tissue.
• At-risk subgroups: None established in humans; this is an in vitro, disease-model, non-clinical finding.
• Counter-evidence: Escitalopram (citalopram's own S-enantiomer) showed the opposite qualitative direction, electron-transport-chain complex I/II+III/IV inhibition, in isolated pig-brain mitochondria at 50-100 µmol/L (Hroudova et al. 2023, cited in the escitalopram dossier in this series). The two findings are not directly comparable: different species/tissue source (mouse neuronal cell line vs. isolated pig brain mitochondria), different disease context (amyloid-stressed vs. unstressed), different endpoints (dynamics/biogenesis/mitophagy markers vs. direct ETC enzyme activity and ATP kinetics), and, critically, citalopram itself was not the molecule tested in the Hroudova study. This divergence is disclosed explicitly per this review's instructions rather than silently reconciled.

Regulatory status

Not addressed in the FDA label.

Population evidence

Reddy AP, Yin X, Sawant N, Reddy PH. "Protective effects of antidepressant citalopram against abnormal APP processing and amyloid beta-induced mitochondrial dynamics, biogenesis, mitophagy and synaptic toxicities in Alzheimer's disease." Hum Mol Genet. 2021;30(10):847-864, doi:10.1093/hmg/ddab054, PMID 33615359. Immortalized mouse hippocampal (HT22) cells transfected with mutant APP cDNA, treated with 20 µmol/L citalopram for 48 hours, compared to untreated mutant-APP cells. Reported changes (citalopram-treated vs. untreated mutant-APP cells): Drp1 (fission) 2.5-fold decrease (P=0.004); Fis1 (fission) 2.2-fold decrease (P=0.005); Mfn1 (fusion) 2.0-fold increase (P=0.02); Mfn2 (fusion) 2.7-fold increase (P=0.05); OPA1 (fusion) 2.1-fold increase (P=0.04); PGC1α (biogenesis) 3.1-fold increase (P=0.01); TFAM (biogenesis) 3.1-fold increase (P=0.03); NRF1 2.3-fold increase; NRF2 2.3-fold increase (P=0.004); PINK1 (mitophagy) 2.0-fold increase; BCL2 2.1-fold increase (P=0.05); cell viability increased (P=0.04); Aβ42 levels significantly reduced (P=0.01).

Mechanism

Proposed as citalopram counteracting amyloid-beta-driven mitochondrial fragmentation and biogenesis failure in this specific disease-model cell line; not established as a general effect on normal, non-amyloid-stressed mitochondria, and ATP/membrane-potential were not measured in this study.

Counter-evidence held on record

This is an in vitro, disease-model-specific finding in a single concentration and cell line. Do not present as evidence that citalopram is a general mitochondrial "protectant" in healthy tissue, and do not present it as contradicting or confirming escitalopram's separate in vitro finding; the two rest on different molecules, models, and endpoints (see Overlap map).

Limits on what this shows

• Key negative/null findings: No human oxidative-stress-marker study for citalopram at therapeutic dosing was located in this review (contrast escitalopram's own small human trial, Cimen et al., which found clinical dosing associated with reduced, not increased, systemic oxidative markers, though confounded by depression-state resolution). The citalopram-specific findings here are confined to testicular tissue in mice, not systemic or neural tissue, and not confirmed in humans.
• At-risk subgroups: Not established in humans.
• Counter-evidence / competing explanation: Tissue specificity is a major limiting factor: a testicular-toxicology finding in mice at a dose not confirmed as clinically equivalent should not be generalized to systemic human oxidative-stress risk at standard antidepressant dosing.

Regulatory status

Not addressed in the FDA label.

Population evidence

Moradi M, Hashemian MA, Douhandeh E, et al. "The protective role of melatonin in citalopram-induced reproductive toxicity via modulating nitro-oxidative stress and apoptosis in male mice." 2023, PMID 36966901. Moradi M, Hashemian MA, Fathi M, et al. "Utility of vitamin C in ameliorating citalopram-induced testicular toxicity via modulating nitro-oxidative stress and apoptosis in mice." 2024, PMID 37794744. Both titles confirm a citalopram-induced (not merely associated) nitro-oxidative-stress and apoptosis effect in male mouse testicular tissue, each ameliorated by a co-administered antioxidant (melatonin in the first, vitamin C in the second). This review could not retrieve the full abstracts' exact citalopram dose (mg/kg), treatment duration, or the specific magnitude of change in individual markers (e.g., MDA, SOD, catalase, nitric oxide) during this pass; see access notes.

Mechanism

Not fully specified in the material retrieved; titled as "nitro-oxidative stress," implying combined reactive-oxygen and reactive-nitrogen-species involvement, consistent with citalopram's separate, class-level, in vitro MAO-inhibition chemistry (Kosel M et al. 2002, PMID 11840311, citalopram biotransformation by monoamine oxidase) as a plausible but not independently confirmed contributing pathway.

Counter-evidence held on record

Do not present melatonin or vitamin C as a citalopram "antidote" or as evidence that citalopram routinely depletes either compound in humans; the cited studies test their protective co-administration against a specific mouse testicular-toxicology endpoint, not human depletion of melatonin or vitamin C by citalopram. No supplement dose is recommended on this basis (see Safety review and Scope and limits).

Limits on what this shows

• Key negative/null findings: No study was located measuring serum iron, ferritin, or transferrin saturation in citalopram-treated patients. This entry is a mechanism/outcome-adjacent finding, not a measured nutrient finding, so it cannot be promoted above yellow regardless of how well-powered the bleeding-odds-ratio meta-analysis is.
• At-risk subgroups: Concurrent NSAID, aspirin, or anticoagulant use substantially raises class-level GI bleeding risk (a well-documented interaction in its own right, though not iron-specific).
• Counter-evidence: None located specific to citalopram and iron status either way; the gap between a real bleeding-outcome signal and any actual iron-status measurement is the main limiting factor, exactly as it was for escitalopram, though the underlying bleeding-risk evidence is now drug-specific rather than class-extrapolated.

Regulatory status

Not addressed as an iron-specific warning in the FDA label; bleeding risk generally (including GI bleeding, especially with NSAIDs, aspirin, or anticoagulants) is a labeled class-level warning for SSRIs.

Population evidence

Gomez-Lumbreras A, et al. "Risk of gastrointestinal bleeding by specific SSRIs and SNRIs: A systematic review and meta-analysis." Br J Clin Pharmacol. 2025/2026 (epub ahead of print). 20-study systematic review with drug-specific odds ratios: citalopram 1.38 (95% CI 1.17-1.62), venlafaxine 1.50 (95% CI 1.32-1.70, highest), fluoxetine 1.38 (95% CI 1.26-1.51), paroxetine 1.31 (95% CI 1.07-1.62). Citalopram was directly named with its own drug-specific estimate in this analysis, unlike escitalopram, which was not retrievable among the drug-specific estimates in the secondary coverage opened for this review (see the escitalopram dossier in this series).

Mechanism

Named and pharmacologically direct: platelets lack de novo serotonin synthesis and depend on plasma reuptake via the same serotonin transporter citalopram inhibits therapeutically; chronic depletion of platelet-dense-granule serotonin impairs serotonin-potentiated platelet aggregation.

Counter-evidence held on record

Do not recommend iron supplementation on this basis. GI bleeding is a clinical outcome, not iron deficiency in itself; iron depletion would only follow from chronic or repeated blood loss, which has not been demonstrated for citalopram specifically via any iron-status measurement. This entry is theoretical and outcome-based, not a confirmed nutrient deficit.

Limits on what this shows

• Key negative/null findings: Brain (whole-brain, hippocampal, cerebellar as separately assayed) copper was explicitly unchanged by chronic citalopram in this study, a genuine drug-specific null finding, in contrast to ECS, which did raise hippocampal and cerebellar copper.
• At-risk subgroups: Not established; rat model only, no human copper data for citalopram were located.
• Counter-evidence: This entry is itself substantially counter-evidence against any copper-depletion claim for citalopram at the brain-tissue level in this animal model.

Regulatory status

Not addressed in the FDA label.

Population evidence

Schlegel-Zawadzka M, Nowak G. "Alterations in serum and brain trace element levels after antidepressant treatment. Part II. Copper." Biol Trace Elem Res. 2000 Jan;73(1):37-45, PMID 10949967. Companion paper (Nowak G, Schlegel-Zawadzka M, "...Part I. Zinc," Biol Trace Elem Res. 1999;67(1):85-92, PMID 10065601) confirms the same rat cohort design compared chronic citalopram, chronic imipramine, and electroconvulsive shock (ECS) arms. Confirmed direct quote from the Part II (copper) abstract: "Chronic treatment with both drugs did not alter the brain copper level. However, chronic ECS induced a significant increase (by 36%) in the copper level in the hippocampus and also in the cerebellum (by 16%)." "Both drugs" refers to citalopram and imipramine, per the companion Part I paper's drug list. This review could not confirm the paper's reported serum-copper direction/magnitude for citalopram specifically after repeated attempts (see access notes); no serum-copper number for citalopram is asserted here without that confirmation.

Mechanism

Not established; this is a descriptive rat pharmacology finding, not a mechanistic study.

Counter-evidence held on record

Present as "a rat study found no change in brain copper with chronic citalopram treatment," not as evidence of any copper effect, deficiency or excess, in humans. No supplement or monitoring recommendation follows from this null animal finding.

Limits on what this shows

• Key negative/null findings: Male offspring were explicitly reported as unaffected in the same study, per its own title, a genuine sex-specific null finding within the same dataset.
• At-risk subgroups: This is a developmental/perinatal-exposure finding in rats; it does not establish an effect in adult humans taking citalopram for depression, which is a materially different exposure window and species.
• Counter-evidence: No directly contradicting citalopram-specific study located. The evidence base itself (single animal study, developmental exposure only) is the main limiting factor, not a competing dataset.

Regulatory status

Not addressed in the FDA label.

Population evidence

Kropp DR, Glover ME, Samanta R, Unroe KA, Clinton SM, Hodes GE. "Perinatal citalopram exposure alters the gut composition and microbial metabolic profiles of Sprague-Dawley rat dams and female offspring but not male offspring." 2025, PMID 41340151. Title-level and PubMed-search-listing-confirmed findings: Sprague-Dawley rat model, perinatal (pregnancy/lactation-window) citalopram exposure, sex-divergent effect (dams and female offspring affected, male offspring not). This dossier could not retrieve the full abstract's exposure dose, exact taxa, or specific metabolic-pathway findings during this review; see access notes.

Mechanism

Not established in the material retrieved for this review. Broader gut-brain-axis literature (class-level, not independently re-verified drug-by-drug here) proposes serotonergic signaling and direct antimicrobial-like SSRI activity against some gut bacteria in vitro; disclosed as class-level context only.

Counter-evidence held on record

Present as "a rat study found perinatal citalopram exposure altered gut microbiome composition in dams and female offspring, not male offspring," never as evidence of a gut-microbiome effect in adult humans taking citalopram for depression. No probiotic or supplement recommendation follows from this single, narrowly-scoped animal dataset.

Limits on what this shows

• Key negative/null findings: A second citalopram-specific hepatotoxicity study surfaced in the same search pass, Ahmadian E, et al. "In vitro and in vivo evaluation of the mechanisms of citalopram-induced hepatotoxicity," 2017, PMID 27271269, independently confirms, via a directly quoted PubMed search-listing snippet obtained in this review ("citalopram (20 mg/kg)-induced damage. citalopram-induced oxidative stress cytotoxicity markers were significantly prevented by antio[xidant]..."), that citalopram itself produces oxidative-stress-mediated hepatocyte cytotoxicity in vivo and in vitro. This lends plausibility to a real citalopram-hepatocyte-energetics effect rather than an artifact of the Abegg et al. comparator design, but this second study's own abstract does not itself name an ATP measurement in the text retrieved, so it is disclosed here only as corroborating context, not as independent ATP-specific confirmation, and it is not otherwise cited elsewhere in this dossier.
• At-risk subgroups: Not established in humans; both located studies are in vitro (and, for Ahmadian et al., additionally in vivo rat) toxicology screens, and the Abegg et al. ATP finding is a hepatocyte-specific, not systemic, endpoint.
• Counter-evidence: None located contradicting an ATP-depletion effect specifically; the main limiting factor is confirmation depth (search-listing/title level only, despite genuine further attempts across five additional channels) rather than a competing negative dataset. Escitalopram's own supraphysiologic electron-transport-chain-inhibition data (Hroudova et al. 2023) remain inapplicable here, since citalopram itself was not the molecule tested in that study; this new finding is citalopram-specific and evaluated independently of the escitalopram dataset, not as a substitute for it.

Regulatory status

Not addressed in the FDA label.

Population evidence

Abegg VF, et al. "Mechanisms of hepatocellular toxicity associated with the components of St. John's Wort extract hypericin and hyperforin in HepG2 and HepaRG cells." 2024, PMID 38219807. Confirmed via a direct fetch of the PubMed search-listing page for this record: this in vitro toxicology study in human HepG2 and HepaRG hepatocyte-derived cell lines measured ATP depletion as a cytotoxicity endpoint across a panel of reference/comparator compounds that includes citalopram, alongside the St. John's Wort constituents hypericin and hyperforin. Despite an additional, dedicated multi-channel attempt in this review (direct PubMed article page, Europe PMC REST API, Semantic Scholar Graph API, Crossref API, and NCBI eutils, each tried at least once with appropriate pacing between calls), this dossier could not retrieve the full abstract text to confirm citalopram's exact tested concentration, exposure duration, or ATP-depletion magnitude relative to the other compounds; every one of those additional channels returned a bot-verification interstitial, an HTTP 429 rate-limit response, or (for eutils) a robots.txt disallowal during this pass. The finding is confirmed at the PubMed search-listing/title level, the same confirmation standard already used elsewhere in this dossier for the Kropp et al. and Moradi et al. entries.

Mechanism

Not established in the material retrieved. ATP depletion in a hepatocyte cytotoxicity screen of this kind is typically read as a marker of mitochondrial or broader cellular energetic compromise at the in vitro concentrations tested, not necessarily indicative of an effect operating at therapeutic human plasma concentrations.

Counter-evidence held on record

This is a liver-hepatocyte in vitro cytotoxicity-screening finding, not a measurement of ATP production in neurons, muscle, or the body generally, and not established at therapeutic human plasma concentrations. Do not present this as evidence that citalopram causes clinically meaningful energy-metabolism impairment or fatigue in patients on standard antidepressant dosing. No supplement, CoQ10, or monitoring recommendation follows from this single, tissue-specific, note-confirmed finding.

Limits on what this shows

• Why no grade: Dual condition met, now with the previously untried erythrocyte/red-blood-cell angle completed and with an additional dedicated, unshared research-tool budget applied specifically to close this gap. Note explicitly: citalopram's labeled QTc-prolongation and dose-cap warning (see Pharmacokinetic basis) is a cardiac-electrical-conduction risk, not a hemoglobin, hematocrit, red-cell-mass, or tissue-oxygen-delivery finding, and is not counted toward this category per Hard Rule 2 (a drug's own intended or labeled pharmacologic risk is not a nutrient-style finding); it is addressed instead in the Monitoring panel.

Channels tried

From the prior review: PubMed search-listing phrasings "citalopram hemoglobin hematocrit anemia" and "citalopram oxygen consumption tissue hypoxia" (both 0 results, clean). In this repair pass: the previously untried "citalopram erythrocyte / red blood cell" angle was completed via the Crossref works-search API ("citalopram erythrocyte red blood cell hemoglobin"), which returned a clean, non-rate-limited result set of 15 works spanning 1975-2026 on red-cell structure, hemoglobin-based oxygen carriers, malaria/parasitized erythrocytes, and blood substitutes, none of which studied citalopram or any SSRI. A further angle, tissue oxygenation and cytochrome c oxidase activity ("citalopram tissue oxygenation cytochrome oxidase"), was attempted via PubMed's search-listing page but was blocked by proxy-level rate-limiting (HTTP 429) after repeated, well-spaced retries. Direct pubmed.ncbi.nlm.nih.gov article pages, the Europe PMC REST API, the Semantic Scholar Graph API, the OpenAlex works API, and a second Crossref query (for the ATP category, sharing the same channel budget) were also attempted in this pass and were blocked by bot-verification interstitials, target-side HTTP 429 responses, or proxy-level rate-limit holds instructing no further retries on that host this session.

Assessment

Three of four attempted phrasings across this review and the prior one returned clean, non-rate-limited zero results, including the erythrocyte/red-blood-cell angle specifically flagged as untried in the prior pass, now completed via Crossref. Despite a dedicated, unshared research-tool budget applied specifically to this category in this repair pass, the remaining channels (PubMed direct pages, Europe PMC, Semantic Scholar, OpenAlex, a second Crossref query) were blocked by the same class of host-level and proxy-level limits encountered in the original review, indicating this reflects infrastructure-level access constraints at the time of both reviews rather than a gap specific to shared-session load. Given the consistent pattern of clean zero-result searches across every phrasing that did complete, this is now assessed as more likely a genuine literature gap than a tooling artifact, though a future revision should still retry the cytochrome-oxidase/tissue-oxygenation angle and the blocked API channels directly, ideally at a time of day or from infrastructure with more headroom against these hosts' own rate limits.