Hydroxyzine

Limits on what this shows

• Null/negative findings: None identified in this or any other tissue; no study was located measuring mitochondrial parameters in normal (non-cancer) human cells at clinical hydroxyzine exposure.
• At-risk subgroups: None established. This is an in vitro, supratherapeutic-concentration finding in malignant cells; no in vivo or clinical-dose correlate was located.

Regulatory status

Not addressed in FDA or EMA-aligned labeling; no mitochondrial-toxicity warning.

Population evidence

Shakya R, Park GH, Joo SH, Shim JH, Choi JS. Hydroxyzine Induces Cell Death in Triple-Negative Breast Cancer Cells via Mitochondrial Superoxide and Modulation of Jak2/STAT3 Signaling. Biomol Ther (Seoul). 2022. In vitro, human BT-20 and HCC-70 triple-negative breast cancer cell lines, hydroxyzine concentrations of 0, 5, 10, 20, 50, 75, and 100 micromolar over 24, 48, and 72 hours: dose- and time-dependent apoptosis with activation of caspase-3, -7, -8, and -9 and cleaved PARP; mitochondrial superoxide generation confirmed by MitoSOX Red staining; apoptosis was prevented by pre-treatment with the antioxidant N-acetylcysteine or with Mito-TEMPO, a mitochondria-targeted antioxidant, confirming the effect is mitochondrial-ROS mediated. Mechanistically linked to down-regulation of JAK2/STAT3 phosphorylation.

Mechanism

Mitochondrial superoxide/reactive-oxygen-species generation driving JAK2/STAT3-linked apoptotic signaling in malignant cells; a cytotoxicity mechanism the authors frame for drug-repurposing potential, not a mechanism described as operating in normal tissue at standard antihistamine or anxiolytic dosing.

Counter-evidence held on record

The source study did not report a direct comparison to clinical plasma levels, and this review could not independently retrieve a verified hydroxyzine Cmax figure to construct one this session (a tooling limitation, disclosed here rather than asserting an unverified number). The concentrations tested, 5 to 100 micromolar, were chosen to kill cancer cells and are far above what oral antihistamine or anxiolytic dosing is expected to produce. This does not establish mitochondrial harm at standard therapeutic dosing in normal tissue and should not be read as a basis for mitochondrial-support supplementation, for which no dosing evidence exists in this drug's population. See Overlap map: this source is shared with the Oxidative stress entry below.

Limits on what this shows

• Null/negative findings: None identified.
• At-risk subgroups: None established in humans; the only direct evidence is in vitro cancer cells and an access-limited animal study.

Regulatory status

Not addressed in FDA or EMA-aligned labeling.

Population evidence

Shakya et al. 2022 (as above): mitochondrial superoxide/ROS generation confirmed via MitoSOX Red, rescued by N-acetylcysteine and Mito-TEMPO, at 5-100 micromolar in breast cancer cell lines. Megdad-Lamraoui A, et al. Cerebrospinal inflammatory response following scorpion envenomation: role of histamine H1 and H3 receptors. 2019. In vivo, hydroxyzine administered directly at 10 mg/kg to test the role of histamine H1 and H3 receptor blockade in a scorpion-envenomation inflammatory model; the study is confirmed to have tested hydroxyzine specifically at the title and bibliographic-record level, but the full abstract text could not be retrieved this session (PubMed presented a bot-verification challenge page and Europe PMC returned repeated HTTP 429 responses on every attempt), so the direction and magnitude of any oxidative-stress or inflammatory-marker change could not be independently verified and is not used to support this grade's direction.

Mechanism

Mitochondrial superoxide/ROS generation is directly demonstrated for hydroxyzine in the Shakya cancer-cell model. No systemic, whole-body direction (pro-oxidant or antioxidant) is established at clinical dosing.

Counter-evidence held on record

Do not read this as an established systemic pro-oxidant or antioxidant effect at ordinary clinical dosing. The Shakya finding is confined to a cancer-cell cytotoxicity model at high concentration, and the Megdad-Lamraoui finding could not be verified beyond its title this session. See Overlap map: the Shakya source is shared with Mitochondrial burden above.

Limits on what this shows

• Why no grade: Both conditions for ungraded are met, now on firmer footing than the original review: one genuine, completed, drug-specific literature search found nothing, the FDA label is silent on copper, and the consumer sources checked show no circulating claim. The remaining gap is that every attempted class-level (Tier C) extrapolation search, which could have justified a yellow grade if a related first-generation antihistamine had documented copper data, was blocked before completing, so a class-extrapolation basis can be neither confirmed nor ruled out. A further revision should specifically retry the antihistamine-class copper queries when Europe PMC, Crossref, OpenAlex, and Semantic Scholar are not policy-blocking requests, before this gap is treated as fully closed.

Channels tried

This pass: the Europe PMC REST search API ("hydroxyzine AND copper"), the Crossref works API ("hydroxyzine copper trace mineral"), the OpenAlex works-search API (two phrasings), the Semantic Scholar Graph API, PubMed search-listing pages, and, for class-level extrapolation, Europe PMC and Crossref queries combining copper with "antihistamine," "cyproheptadine," and "diphenhydramine." The FDA label was re-checked directly for any copper-related adverse-effect language. Consumer sources re-checked directly for copper specifically: the Drugs.com hydroxyzine side-effects page and a Suzy Cohen site search for "hydroxyzine copper." (Prior pass also checked a public prescription drug-nutrient depletion chart PDF, not re-checked this pass since it does not list hydroxyzine at all.)

Outcome

The Europe PMC drug-specific query completed cleanly and returned 25 results, none relating hydroxyzine to copper, ceruloplasmin, or trace-mineral status; results were miscellaneous publications matching "hydroxyzine" or "copper" as separate keywords (liver disease and sleep, pediatric dental sedation, drug repurposing, sarcoidosis, and similar, unrelated to copper metabolism). Every other attempt, including both class-level extrapolation queries, returned either a repeated HTTP 429 response from the agent proxy or, on PubMed, a reCAPTCHA bot-verification page, and did not complete. The proxy's own status endpoint confirmed these as HTTP 403 "policy denial or upstream failure" rejections at the gateway level for pubmed.ncbi.nlm.nih.gov, api.crossref.org, api.openalex.org, and api.semanticscholar.org during this session, consistent with an organization-level access policy rather than ordinary rate-limiting. This session's general web search tool reported its budget already exhausted (200 of 200) before any search for this dossier could run. The FDA label contains no mention of copper, iron, anemia, hemoglobin, hematocrit, or hematologic adverse effects at all. Both consumer sources returned no hydroxyzine-copper claim (Drugs.com: no match; Suzy Cohen: "0 Search Results Found" for "hydroxyzine copper").

Limits on what this shows

• Why no grade: Both conditions for ungraded are met, and confidence is now higher than in the original review: two independent, cleanly completed searches on two different hosts (PubMed and Europe PMC), across two separate research passes, found nothing supporting even a yellow grade, the FDA label is silent on iron and hematologic effects, and no circulating claim exists to rebut. The remaining gap is the blocked class-level (anticholinergic-and-iron-absorption) extrapolation query, which a further revision should retry when Crossref is not policy-blocking requests.

Channels tried

Original pass: PubMed search-listing pages ("hydroxyzine iron", "hydroxyzine AND iron AND status") and the Europe PMC REST search API. This follow-up pass: the Europe PMC REST search API again ("hydroxyzine AND iron"), the OpenAlex works-search API ("hydroxyzine iron hematologic"), and the Crossref works API (an anticholinergic-and-iron-absorption class-extrapolation query). The FDA label was re-checked directly for iron- or hematologic-related adverse-effect language, and the Drugs.com side-effects page was re-checked directly.

Outcome

Original pass: the "hydroxyzine iron" PubMed phrasing completed and returned five results, none relevant (a polycythemia vera pruritus paper, a potentially-inappropriate-medications list naming both drugs without linking them causally, a cetirizine-diclofenac interaction paper, a restless legs syndrome case report, and a histamine-CYP450 liver cell study); the narrower phrasing and the Europe PMC attempt both returned HTTP 429. This follow-up pass: the Europe PMC query completed cleanly and independently, returning 25 results, none relevant (sickle cell pain crises, epidermolysis bullosa pain management, cholestasis and pruritus, iron nanoparticles in unrelated periodontal research, and similar keyword collisions, none examining hydroxyzine's effect on iron). The OpenAlex attempt returned a 503 server error (a target-side fault, not a policy block) and the Crossref class-extrapolation query was rejected by the agent proxy with HTTP 429/403 before completing. The FDA label contains no mention of iron, anemia, hemoglobin, hematocrit, or hematologic adverse effects. Drugs.com returned no hydroxyzine-iron claim.

Limits on what this shows

• Null/negative findings: None identified for hydroxyzine specifically; the general absence of any hydroxyzine-microbiome literature located in this review is itself the primary finding for this entry.
• At-risk subgroups: None established.

Regulatory status

Not addressed in FDA or EMA-aligned labeling.

Population evidence

Hammouda ZK, Wasfi R, Abdeltawab NF. Fexofenadine HCl enhances growth, biofilm, and lactic acid production of Limosilactobacillus reuteri and Bifidobacterium longum: implications for allergy treatment. BMC Microbiol. 2025. In vitro, tested fexofenadine, cyproheptadine, and desloratadine against cultures of two beneficial gut bacterial species: fexofenadine enhanced growth, biofilm formation, and lactic and acetic acid production; per the indexed abstract, this beneficial effect was seen with fexofenadine "unlike cyproheptadine and desloratadine." Hydroxyzine itself was not in this study's cohort, a class-extrapolation limitation under Hard Rule 6. Cyproheptadine, the tested drug that most resembles hydroxyzine pharmacologically (also first-generation, sedating, and strongly anticholinergic, unlike fexofenadine and desloratadine, which are non-sedating, minimally anticholinergic second-generation agents), is the one that did not show the fexofenadine-associated benefit; the exact direction of cyproheptadine's own effect (neutral versus inhibitory) could not be independently confirmed beyond the abstract-level record accessed this session.

Mechanism

Not established for hydroxyzine directly. Anticholinergic effects on gastrointestinal motility and secretions offer general biological plausibility for a gut-microbiome effect, in either direction, but no hydroxyzine-specific data confirm it.

Counter-evidence held on record

This is class extrapolation from a study that did not include hydroxyzine, and the pharmacologically closest analog tested did not show the beneficial effect seen with a dissimilar, non-anticholinergic class member. This should not be read as evidence that hydroxyzine either harms or helps the gut microbiome; it is a disclosed absence of drug-specific data with weak, direction-ambiguous class-level context.

Limits on what this shows

• Null/negative findings: This entire entry is a null finding. It is listed here, rather than only in the counter-evidence record, because the protocol treats a genuine drug-specific null result as a valuable, disclosable finding in its own right, and because it is the only direct hydroxyzine-ATP data located.
• At-risk subgroups: None established.

Regulatory status

Not addressed in FDA or EMA-aligned labeling.

Population evidence

Wilson WS. The effects of phenobarbitone, leptazol, dexamphetamine, iproniazid, imipramine, LSD, chlorpromazine, reserpine and hydroxyzine on the in vivo levels of adenine nucleotides and phosphocreatine in the rat brain. Br J Pharmacol. 1969. In vivo rat brain, enzymic and ion-exchange chromatographic measurement of ATP, ADP, AMP, and phosphocreatine after administration of several psychotropic drugs: hydroxyzine, along with dexamphetamine and LSD, "showed no significant changes" in these parameters at the doses tested. By contrast, in the same study, an anaesthetic dose of phenobarbitone increased ATP and phosphocreatine while reducing ADP and AMP, the convulsant leptazol produced the opposite pattern, and reserpine decreased the ATP/ADP ratio, indicating the assay was sensitive enough to detect a real effect had one been present at the hydroxyzine dose tested.

Mechanism

None demonstrated; this entry rests on a direct negative finding, not a proposed mechanism.

Counter-evidence held on record

A single, roughly 55-year-old rat study using older biochemical assay methods, at an unspecified and unreplicated dose, is not a substitute for modern human data, and no modern replication was located. This finding is reassuring rather than concerning, but it does not rule out an effect at other doses, exposure durations, or in other tissues. It is unrelated to the in vitro mitochondrial-ROS finding described under Mitochondrial burden; see Overlap map, which records that no evidence overlap exists between these two entries.

Limits on what this shows

• Why no grade: Both conditions for ungraded are met, now with two independent completed searches (the original general web search and this pass's Crossref query) rather than one, plus a direct, negative re-check of the FDA label's hematologic content. No circulating claim was found on the sources checked. The remaining gap is the blocked Europe PMC and class-extrapolation queries; a further revision should retry those specifically when Europe PMC and Crossref are not policy-blocking requests.

Channels tried

Original pass: a general web search for hydroxyzine plus hemoglobin, hematocrit, and oxygen saturation, and a PubMed search-listing query combining those terms. This follow-up pass: the Crossref works API ("hydroxyzine oxygen hemoglobin"), the Europe PMC REST search API (two phrasings: hemoglobin/hematocrit/oxygen-saturation combined, and "oxygen delivery" alone), and, for class-level extrapolation, a Crossref query combining "H1 antihistamine" with hematocrit and red blood cell terms. The FDA label was re-checked directly for hemoglobin, hematocrit, anemia, or red-blood-cell language. Consumer sources checked directly across both passes: Drugs.com hydroxyzine side-effects page and Suzy Cohen site search.

Outcome

Original pass: the general web search returned only generic oxygen-saturation and hematocrit physiology pages unrelated to hydroxyzine. This follow-up pass: the Crossref query completed cleanly and returned 10 results, all concerning hemoglobin-based oxygen carriers and blood-substitute research (regulatory frameworks for HBOC trials, ischemic rescue, hemoglobin biochemistry, artificial-cell nanomedicine), a pure keyword collision on "hemoglobin" and "oxygen" with no hydroxyzine relevance. Both Europe PMC phrasings and the Crossref class-extrapolation query returned repeated HTTP 429/403 responses from the agent proxy and did not complete; PubMed again presented a reCAPTCHA bot-verification page. This session's general web search tool could not be used for further queries, its 200-call budget already exhausted before this dossier's research began. The FDA label's only hematologic-adjacent adverse-effect entry is a postmarketing report of thrombocytopenia (a platelet/clotting effect); it contains no mention of hemoglobin, hematocrit, anemia, or red-blood-cell count, so thrombocytopenia is noted here for completeness but correctly excluded from this category, which concerns oxygen-carrying capacity and delivery, not platelets. Hydroxyzine's labeled QT-interval prolongation and Torsade de Pointes risk (see Pharmacokinetic basis and the Potassium and magnesium entry above) is a cardiac electrophysiology risk, not a measure of oxygen delivery, and remains deliberately excluded from this category under Hard Rule 2. Both consumer sources returned no hydroxyzine-oxygen-delivery claim.