Perphenazine

Limits on what this shows

• Key negative/null findings: The headline result in healthy-donor cells is itself a null finding: basal, ATP-linked, maximal, and reserve-capacity OCR, proton leak, and non-mitochondrial respiration were all unchanged. This argues against a universal mitochondrial-toxicity claim for perphenazine and is recorded here as the dossier's own counter-evidence.
• At-risk subgroups: The altered signal was specific to lymphocytes derived from unmedicated patients with schizophrenia or bipolar disorder, not from healthy donors. Whether this reflects a pre-existing illness-related mitochondrial vulnerability unmasked by drug exposure, rather than a direct perphenazine effect, is not established by this ex vivo design.
• Counter-evidence on record: The healthy-donor null result above. Also note this is an ex vivo lymphocyte assay, not an in vivo clinical mitochondrial-function outcome, and perphenazine's own short-term clinical-treatment arm in this study had only one patient per diagnostic group, too small to interpret.

Regulatory status

Not addressed in the FDA label; no mitochondrial-toxicity term is listed.

Population evidence

Bar-Yosef T, Hussein W, Yitzhaki O, et al. "Mitochondrial function parameters as a tool for tailored drug treatment of an individual with psychosis." Sci Rep. 2020;10:12258. PMID 32703977. Fresh human lymphocytes (Ficoll-Paque isolation) from healthy controls (n=40), schizophrenia patients (n=48), and bipolar disorder patients (n=27) were exposed in vitro to therapeutically relevant concentrations of several antipsychotics and mood stabilizers, including perphenazine, for 24 hours. Six respiration parameters (basal, ATP-linked, maximal, and reserve-capacity oxygen consumption rate (OCR), proton leak, and non-mitochondrial respiration) were measured on a Seahorse XF-24 analyzer; one-way ANOVA with Fisher's LSD post hoc, p<0.05. In healthy-donor-derived cells, perphenazine (like valproic acid) did not significantly change any of the six parameters. In cells derived from untreated patients, perphenazine's effect on proton leak and on basal, maximal, and reserve-capacity OCR differed significantly from its effect in healthy-donor cells, and it significantly decreased levels of the mitochondrial fusion protein MFN1 specifically in the patient-derived, perphenazine-exposed condition.

Mechanism

Named as an illness-state-dependent alteration in mitochondrial fission-fusion balance (reduced MFN1) and respiratory-capacity parameters; not established as a direct, universal mitochondrial toxin, since the same exposure produced no measurable change in normal (healthy-donor) cells.

Counter-evidence held on record

This is an ex vivo cell-culture finding, not a measured clinical outcome, and it is bounded to illness-derived cells. No mitochondrial-support supplement (such as CoQ10 or L-carnitine) is indicated by this evidence. Shares its entire evidentiary source with the ATP production entry below; see Overlap map.

Limits on what this shows

• Key negative/null findings: Protein carbonyl, one of the five markers measured, showed no significant difference between the three drug groups (p=0.8), an explicit null result on record.
• At-risk subgroups: Not defined beyond the whole perphenazine-treated cohort in this study; no dose-response or treatment-duration stratification was reported.
• Counter-evidence on record: This is an active-comparator, cross-sectional study with no drug-naive or healthy-control arm (per this protocol's standing limitation on active-comparator designs), so it cannot distinguish a perphenazine-caused deficit from a clozapine-specific antioxidant advantage or from confounding by indication/illness severity. The protein carbonyl null result (above) also tempers the finding.

Regulatory status

Not addressed in the FDA label.

Population evidence

Hendouei N, Farnia S, Mohseni F, et al. "Alterations in oxidative stress markers and its correlation with clinical findings in schizophrenic patients consuming perphenazine, clozapine and risperidone." Biomed Pharmacother. 2018;104:699-704. PMID 29710513. DOI 10.1016/j.biopha.2018.04.109. Descriptive study, n=100 chronic schizophrenia patients (minimum 3 months on current medication, at least 2 years of illness), divided into perphenazine, clozapine, and risperidone groups. Serum glutathione (GSH), protein carbonyl, lipid peroxidation/malondialdehyde (MDA), superoxide dismutase (SOD), and ferric reducing ability of plasma (FRAP) were measured. The clozapine group had significantly higher SOD and lower lipid peroxidation than the other two groups. In the perphenazine and risperidone groups, MDA correlated positively with PANSS positive, negative, and general symptom subscales (a poorer antioxidant profile associated with more severe symptoms); in the clozapine group this correlation was negative. Significance threshold p<0.05.

Mechanism

Not causally established by this design. The authors associate oxidative imbalance with symptom severity; whether perphenazine itself drives the oxidative-stress markers, whether clozapine is unusually antioxidant-favorable, or whether illness severity/chronicity confounds the comparison is not resolved.

Counter-evidence held on record

This finding compares perphenazine against other actively prescribed antipsychotics in already-ill, chronically treated patients, not against no treatment. No antioxidant supplement or dose is recommended on the basis of this evidence.

Limits on what this shows

• Key negative/null findings: None found; no later replication or rebuttal of the 1956 finding was located in the channels searched (PubMed "perphenazine copper ceruloplasmin", zero results; PubMed "chlorpromazine copper", 63 results, this being the one clinically relevant hit on the first page).
• At-risk subgroups: None defined; a single 1956 case-level report with no stated sample size.
• Counter-evidence on record: None found. Perphenazine's own pharmacokinetic profile (extensive CYP2D6 metabolism; no cholangiolitic hepatotoxicity signal identified in the label text retrieved for this review) was not itself tested for this specific copper effect, so this entry cannot confirm or rule out whether perphenazine shares chlorpromazine's pattern.

Regulatory status

Not addressed in the FDA perphenazine label.

Population evidence

Class member with real data, perphenazine absent from cohort: Azima H, Richman A. "Alteration of copper metabolism in chlorpromazine-treated cases." Arch Neurol Psychiatry. 1956;75(2):163. PMID 13282535. Clinical case-level report describing chlorpromazine-treated patients with copper-metabolism alterations resembling Wilson's disease (hepatolenticular pattern), proposed as an early-stage manifestation related to the drug's cholangiolitic (not parenchymal) hepatotoxicity. Chlorpromazine is a phenothiazine, the same core class as perphenazine, but has an aliphatic side chain rather than perphenazine's piperazine side chain, a real chemical difference this review could not resolve the relevance of.

Mechanism

Proposed hepatolenticular/Wilson's-disease-like copper retention linked to cholangiolitic liver injury from chlorpromazine specifically; no mechanism tied to perphenazine's own piperazine side-chain chemistry was located.

Counter-evidence held on record

This is a reverse-direction, class-extrapolated (Tier C) signal pointing toward copper retention, not deficiency, from a single, unreplicated, 70-year-old report in a different phenothiazine. No copper supplementation is implied or supported by this evidence; if this class pattern did extend to perphenazine, it would be a theoretical reason for caution with added copper, not a reason to supplement it.

Limits on what this shows

• Why no grade: Both sides of the dual-condition test point to ungraded, with a disclosed caveat. No perphenazine-specific iron, ferritin, or hematologic study was found: PubMed "perphenazine iron" returned exactly 3 results, all unrelated analytical-chemistry assay-development papers (spectrophotometric and molecularly-imprinted-polymer detection methods using iron compounds as reagents), none clinical. A well-known general literature thread exists associating low serum iron with neuroleptic malignant syndrome across "neuroleptics" broadly (Rosebush PI, Mazurek MF. "Serum iron and neuroleptic malignant syndrome." Lancet. 1991;338:149-151, PMID 1677067, plus related reports by Lee 1998, Weller and Kornhuber 1993, and Patil et al. 2014, alongside a contemporaneous rebuttal/null thread including Gurrera 1991), but this review could not confirm which specific antipsychotics, and specifically whether perphenazine or any phenothiazine, were in any of these cohorts. Per this protocol's Hard Rule 6 cohort test, a class label like "neuroleptics" cannot be used to support even a yellow grade until the specific drugs studied are confirmed, so this thread could not be used at any grade level. No circulating consumer claim pairing perphenazine with iron deficiency was identified (checked via the perphenazine Wikipedia article, which names no nutrient, mineral, or hematologic interaction at all).

Channels tried

FDA/DailyMed label (no iron/ferritin term identified in the sections retrieved). PubMed search-listing pages for "perphenazine iron" (3 results, all irrelevant) and "neuroleptic malignant syndrome serum iron" (10 results identified, cohort unconfirmed for all). Direct fetch of the Rosebush and Mazurek 1991 PMID detail page (pubmed.ncbi.nlm.nih.gov/1677067) returned a reCAPTCHA bot-verification challenge on the attempt made. OpenAlex single-record lookup for the same PMID returned title, authors, year, and journal metadata but a null abstract field, no abstract text. Semantic Scholar Graph API lookup for the same PMID returned an HTTP 429 rate-limit response on two separate attempts made after waiting periods. Wikipedia's perphenazine article was checked directly and names no nutrient, mineral, or hematologic interaction.

Genuine gap vs tooling limited

Assessed as tooling-limited rather than a clean literature gap. A real, drug-class-relevant literature thread plausibly exists (the NMS-serum-iron association is well established for neuroleptics generally in the secondary/review literature this review could see titles for), but repeated, independent access blocks across three different channels (a PubMed bot-verification wall, an empty OpenAlex abstract field, and repeated Semantic Scholar rate-limiting) prevented this review from completing the cohort confirmation this protocol requires before using that thread at any grade. This is flagged explicitly for re-check in a future revision with working full-text access to the Lancet/PMC record and the other case reports in this thread.

Limits on what this shows

• Key negative/null findings: None found for perphenazine specifically. Note the chlorpromazine finding is itself in vitro only; no in vivo human gut-microbiome composition study of chlorpromazine or perphenazine was located in this review to confirm clinical relevance.
• At-risk subgroups: None defined; in vitro bacterial-culture evidence only, not a clinical or even animal in vivo finding.
• Counter-evidence on record: None found contradicting the chlorpromazine in vitro result for that drug. No evidence either way was found for perphenazine itself.

Regulatory status

Not addressed in the FDA perphenazine label.

Population evidence

Class member with real data, perphenazine absent from cohort: Seewald M, Behnke JS, Wu H, Wagner MC, Cordes T, Raisch T, Urner LH, Wycisk V. "Detergent Conjugation Decouples Chlorpromazine Solubility from Antibacterial Activity against Bacteria Relevant for Gut Microbiota." Bioconjugate Chemistry. 2026;37(5):881-887. PMID 42001429. In vitro study showing chlorpromazine hydrochloride (the standard clinical salt form) has antibacterial potency against bacteria relevant to gut microbiota composition, correlated with the molecule's cationic, membrane-permeabilizing chemistry; a detergent-conjugated form of chlorpromazine retained water solubility while losing most of this antibacterial potency, which the authors frame as evidence the native drug's antibacterial action, and its potential for microbiota disruption, is a real and separable chemical property.

Mechanism

Cationic amphiphilic, membrane-permeabilizing structure shared by the phenothiazine core that chlorpromazine and perphenazine both have, proposed to drive antibacterial activity; not tested for perphenazine's own piperazine side chain specifically.

Counter-evidence held on record

This is in vitro, class-extrapolated (Tier C) evidence for a different phenothiazine, not a confirmed perphenazine or in vivo human finding. No probiotic, prebiotic, or dietary intervention is recommended on this basis.

Limits on what this shows

• Key negative/null findings: This entry's primary content is itself the null/negative finding, stated as the claim above: no significant change in ATP-linked OCR in healthy-donor cells.
• At-risk subgroups: Whether ATP-linked OCR specifically, as opposed to proton leak or basal/maximal/reserve-capacity OCR, was also differentially affected in the patient-derived cells described under Mitochondrial burden was not reported in the material accessible during this review; this is disclosed as an open question rather than assumed in either direction.
• Counter-evidence on record: None found suggesting impaired ATP production from perphenazine; the available evidence is reassuring.

Regulatory status

Not addressed in the FDA label; no creatine-phosphokinase or energy-metabolism adverse term was identified in the label text retrieved.

Population evidence

Bar-Yosef T, et al. Sci Rep. 2020;10:12258. PMID 32703977 (see Mitochondrial burden entry for full study design, sample sizes, and methodology). ATP-linked oxygen consumption rate was one of six Seahorse-measured respiration parameters; the paper's summary finding that perphenazine "did not affect any mitochondrial respiration parameter" in healthy-donor-derived lymphocytes specifically includes ATP-linked OCR.

Mechanism

None implicated; this is a null finding at a therapeutically relevant concentration in cells from healthy donors.

Counter-evidence held on record

This is not a deficiency finding. No coenzyme Q10, L-carnitine, or other ATP-support supplement is indicated by this evidence. Shares its entire evidentiary source with the Mitochondrial burden entry above; see Overlap map.

Limits on what this shows

• Why no grade: Both sides of the dual-condition test point to ungraded. No perphenazine-specific hematologic or oxygen-delivery study was found in the channels searched. The FDA label sections retrieved for this review (Clinical Pharmacology, Drug Interactions, Adverse Reactions/Endocrine, Precautions) did not surface a hemoglobin, hematocrit, or anemia term, though the full adverse-reaction table was not retrieved and reviewed line by line, a limitation disclosed here rather than treated as a confirmed absence. Agranulocytosis and leukopenia are recognized risks for some phenothiazines as a class, but these are white-cell, intrinsic-toxicity findings, not oxygen-carrying-capacity or nutrient-pathway findings, and are excluded from this category per Hard Rule 2.

Channels tried

PubMed combined query "perphenazine hemoglobin OR anemia OR hematocrit" was attempted but the search-listing page returned results that did not appear to be properly filtered by the perphenazine term (a 338,568-result general hematology page), a disclosed query-parsing failure rather than a true negative; a cleaner single-term retry ("perphenazine" AND "hemoglobin", properly quoted) was attempted but returned an HTTP 429 rate-limit response and could not be completed within this review's remaining channel budget. The Crossref REST API query "perphenazine hemoglobin anemia" returned 10 results, none perphenazine-specific (general sickle-cell, thalassemia, and hemolytic-anemia literature). Wikipedia's perphenazine article was checked directly and names no hematologic or oxygen-delivery interaction; the only metabolic note present is an insulin-needs caution in diabetic patients, which is the drug's own labeled pharmacology, not an oxygen-delivery finding, and is excluded here on that basis.

Genuine gap vs tooling limited

Assessed as mixed. The Crossref channel did return a clean, informative negative result, and the FDA label sections actually retrieved showed no signal, which both lean toward a genuine literature gap for this category. But the PubMed combined-query parsing failure and the rate-limited single-term retry mean the most natural drug-specific search was not cleanly completed, so this is flagged for a cleaner, single-term re-attempt in a future revision rather than treated as fully closed.