Molindone

Limits on what this shows

• Key negative/null findings: Clozapine, also tested in the same experiment, showed markedly weaker Complex I inhibition, tracking its lower clinical EPS liability. Molindone itself is comparatively low-EPS among typical antipsychotics (a pediatric comparator trial, TEOSS, reported less problematic akathisia on molindone than on atypical comparators), which - if the paper's own potency-correlation holds - would suggest, but does not establish, that molindone's own effect (if any) sits toward the lower end of this spectrum rather than the haloperidol end. No study of molindone's own effect on Complex I was located despite search.
• At-risk subgroups: None identified specific to molindone; this is a laboratory, class-level signal only.
• Counter-evidence on record: A separate search for a Burkhardt et al. 1993 (Ann Neurol) report of neuroleptic Complex I inhibition - cited by sibling dossiers in this series for other typical antipsychotics - returned an HTTP 403 block when this review attempted to open the original article directly; it is listed in Sources as found-but-not-opened and was not used as grading evidence here, consistent with not citing a source that was not actually opened in this review.

Regulatory status

Not addressed in the FDA label.

Population evidence

Maurer & Möller 1997 (Mol Cell Biochem) measured Complex I activity in postmortem human brain cortex homogenate exposed in vitro to five neuroleptics: haloperidol (IC50 ≈0.1 mM, most potent), chlorpromazine (≈0.4 mM), risperidone and zotepine (≈0.5 mM each), and clozapine (≈2.5 mM, least potent, roughly 25-fold weaker than haloperidol). Inhibition potency paralleled each drug's known clinical extrapyramidal side-effect burden. Cohort test: haloperidol and chlorpromazine are typical/first-generation antipsychotics sharing molindone's D2-antagonist mechanism (defensible class test per this project's class-extrapolation rule for FGAs); molindone itself was not tested, so this is Tier C for molindone specifically, not Tier A/B.

Mechanism

Named for the class: thiol oxidation at the ubiquinone-binding region of Complex I, reported elsewhere specifically for haloperidol.

Counter-evidence held on record

This is a class-level, not drug-specific, laboratory finding. It does not establish that molindone itself impairs mitochondrial respiration in patients, and no supplementation (CoQ10, carnitine, or otherwise) is supported by it.

Limits on what this shows

• Key negative/null findings: In vivo rat genotoxicity assays were negative at the doses tested; the in vitro mutagenic signal appeared only with metabolic activation and was fully blocked once glutathione/glucuronic acid were added, meaning normal Phase II capacity appears to neutralize the signal under the tested conditions. No human oxidative-stress biomarker (malondialdehyde, superoxide dismutase, glutathione, 8-OHdG) has been measured in molindone-treated patients in any source located for this review.
• At-risk subgroups: Theoretical only: patients with reduced glutathione-conjugation capacity (concurrent high hepatic drug burden, malnutrition, chronic heavy alcohol use) would plausibly buffer this signal less well. This is extrapolated from the mechanism, not from drug-specific outcome data.
• Counter-evidence on record: Krishna et al. 2016's own stated conclusion is reassuring, not alarming: "the weight of evidence from these studies, combined with a large margin of safety and efficient detoxification through Phase II conjugation supports the interpretation that molindone hydrochloride does not pose a genotoxic risk to humans."

Regulatory status

Not addressed. The FDA label states molindone is "rapidly absorbed and metabolized" with 36 recognized metabolites and under 2-3% excreted unchanged, but names no oxidative pathway or redox marker.

Population evidence

Two molindone-specific in vitro/ex vivo studies. (1) Krishna et al. 2016: Ames assay showed up to 3-fold increased revertants in strain TA100, but only with rat liver S9 metabolic activation; this effect was "completely abrogated by the addition of glutathione and UDP-glucuronic acid," indicating Phase II detoxification of a reactive Phase I metabolite. In vivo rat micronucleus and Comet assays were negative. (2) Gawlik et al. 2020: using human liver microsomes and photocatalysis, identified six previously undescribed molindone metabolites formed via hydroxylation, morpholine-ring dealkylation, and dehydrogenation - consistent with oxidative Phase I biotransformation capable of generating reactive intermediates. Neither study measured a human in-vivo oxidative-stress biomarker.

Mechanism

Named: hepatic oxidative (Phase I) biotransformation of molindone generates a reactive/mutagenic intermediate that is conjugated and cleared via glutathione-S-transferase and UDP-glucuronosyltransferase (Phase II) pathways. This is a plausible, mechanistically real, but clinically unconfirmed antioxidant-capacity burden.

Counter-evidence held on record

This is in vitro/ex vivo mechanistic evidence only, not a demonstrated clinical deficiency. It must not be used to justify glutathione, N-acetylcysteine, or other antioxidant supplementation; the cited research does not support that and empiric antioxidant use could complicate interpretation of routine liver-enzyme monitoring.

Limits on what this shows

• Key negative/null findings: The review's own authors state that "prospective short-term data have not supported" the hypothesis that antipsychotic treatment itself causes the iron drop, favoring the competing explanation that lower baseline iron is a pre-existing risk marker for developing akathisia rather than a drug-induced depletion. Overall study quality was rated poor (small, heterogeneous, mostly cross-sectional cohorts, no molindone data).
• At-risk subgroups: None identified specific to molindone; the pooled analysis does not stratify a dihydroindolone or molindone-specific subgroup.
• Counter-evidence on record: Confounding by indication/active-comparator limitation: most pooled studies compared antipsychotic-treated patients with vs. without akathisia, not treated vs. untreated patients, so a drug-causation claim cannot be fully separated from a pre-existing-risk-marker explanation. Akathisia itself is the drug's own extrapyramidal pharmacologic effect (Hard Rule 2) - this entry grades the iron-status association, not akathisia itself, as the nutrient-relevant finding.

Regulatory status

Not addressed in the FDA label; no iron or anemia finding is named for molindone.

Population evidence

Schoretsanitis, Nikolakopoulou, Guinart, Correll & Kane 2020 (Eur Neuropsychopharmacol), a systematic review and meta-analysis of 10 cross-sectional studies (n=395), found serum iron significantly lower in akathisia-positive vs. akathisia-negative antipsychotic-treated patients (SMD -0.49, fixed-effect, p<0.001; -0.55, random-effects, p=0.008). Cohort test: the pooled studies included haloperidol and chlorpromazine (typical/FGA, defensible class test for molindone) alongside multiple second-generation agents (risperidone, paliperidone, quetiapine, aripiprazole, amisulpride, iloperidone, asenapine, brexpiprazole, lurasidone, cariprazine). Molindone was not among the studied drugs; this is Tier C for molindone, not Tier A.

Mechanism

Proposed, not confirmed: iron is a required cofactor for tyrosine hydroxylase and striatal dopamine synthesis/turnover, and low brain iron is separately associated with restless-legs-type movement phenomena, offering a plausible but unproven link to antipsychotic-induced akathisia.

Counter-evidence held on record

This is a cross-sectional association, not evidence that molindone depletes iron. It must not be read as a reason to supplement iron preemptively. A patient who develops akathisia or EPS on molindone should receive a standard anemia/iron-status workup if clinically indicated, not empiric iron supplementation, given iron's own upper-limit toxicity profile.

Limits on what this shows

• Why no grade: Dual condition met: no Tier B/C mechanistic or class-level evidence links molindone, or a defensible typical-antipsychotic class member, to copper or ceruloplasmin status; and no circulating consumer claim specifically alleging copper depletion or excess from molindone or first-generation antipsychotics was located to grade white instead.

Related finding, considered and rejected for cohort

A 2000 case-control study (Herrán et al., Psychiatry Res 94:51-58), already used by several sibling dossiers in this project batch for other phenothiazines, found significantly higher serum copper in 62 schizophrenia outpatients than controls, highest in a 17-patient subgroup on "depot neuroleptics." Re-opened directly via its publisher abstract page for this repair pass: it names no individual depot drug. Molindone has no depot/long-acting-injectable formulation at all, so it cannot be placed in, or defensibly extrapolated from, this unnamed-agent cohort, and it was not carried forward even as Tier C (consistent with how this project's trifluoperazine dossier, which also lacks a depot formulation, treated the same source).

Channels tried

This repair pass: the Crossref works API was reachable (unlike the original pass) and returned clean, on-topic-but-empty result sets for three separate phrasings ("molindone copper ceruloplasmin," "typical antipsychotic serum copper ceruloplasmin," "haloperidol serum copper schizophrenia trace element") - no antipsychotic-and-copper study surfaced in any of them. PubMed, Europe PMC REST, and NCBI eutils/DailyMed were attempted again and were blocked again, this time with the underlying connection itself refused (HTTP 403, policy denial) rather than merely rate-limited, consistent with this project's documented org-level network restriction. OpenAlex's works-search endpoint returned HTTP 429 and was not retried per the fetch tool's own instruction not to re-attempt. This is now a better-confirmed literature gap than the original pass found (multiple working-channel queries returned clean nulls), with a residual, disclosed tooling limitation on the PubMed/Europe PMC/eutils channels specifically.

Limits on what this shows

• Key negative/null findings: The review itself contrasts first- and second-generation mechanisms rather than reporting a null; no molindone-specific or dihydroindolone-specific microbiome or antimicrobial data were located anywhere in this review's own search (Crossref, OpenAlex, direct publisher fetch), and a separate 2022 scoping review (cited below, supporting this entry's limits) found first-generation agents represented only incidentally in the dedicated antipsychotic-microbiome clinical literature.
• At-risk subgroups: None established specific to molindone; this is a class-level laboratory/review-level signal only.
• Counter-evidence on record: A 2022 scoping review (Singh et al., opened in full for this project's original research pass on this drug) found the antipsychotic-microbiome evidence base concentrated in second-generation agents, with first-generation agents appearing in clinical cohorts only incidentally (one 63-patient cohort included a single chlorpromazine-treated patient, not analyzed separately). This tempers, without contradicting, the McDonagh et al. class-level antimicrobial-properties claim: the laboratory/antimicrobial-activity evidence for the FGA class is broader than the clinical in-vivo gut-dysbiosis evidence specific to any first-generation drug.

Regulatory status

Not addressed in the FDA or DailyMed label.

Population evidence

McDonagh, Murray, Hallahan & Miliotis, J Psychiatr Res 2025 (PMID 40499249): a systematic review of 26 studies on antipsychotic antimicrobial properties, 4 on microbiome alterations, and 31 on infection risk. Its abstract states: "First-generation antipsychotics were observed to have broad antimicrobial properties, whereas second-generation antipsychotics primarily affected commensal bacteria," and that "at least four antipsychotics were observed to disrupt the gut microbiota." Cohort test: this review classifies its antimicrobial finding at the first-generation/typical-antipsychotic class level, which molindone unambiguously belongs to, so it is a defensible class extrapolation (Tier C) per this project's rule for typical/FGA-class evidence; the abstract text accessible in this review does not itself name which individual first-generation agents (e.g., haloperidol, chlorpromazine, molindone) were among the 26 antimicrobial-property studies, and the full text sits behind a publisher paywall this review could not open, so that drug-level detail is disclosed as unconfirmed rather than assumed.

Mechanism

Named for the class only: direct antibacterial activity of first-generation antipsychotics (a property independently documented for individual phenothiazines such as chlorpromazine), proposed by the review's authors as a driver of gut-microbiota disruption distinct from the metabolic/weight-gain-mediated pathway more associated with second-generation agents.

Counter-evidence held on record

This is a class-level, not molindone-specific, finding, and the accessible portion of its source does not confirm which first-generation drugs were actually tested. It is not evidence that molindone itself disrupts the gut microbiome, and it is not a basis for probiotic or other supplementation advice. This source is also used for class-level Gut microbiome risk extrapolation in at least one sibling dossier in this same project batch (thiothixene); see Overlap map.

Limits on what this shows

• Key negative/null findings: No study measured ATP yield or cellular energy charge directly in neuroleptic-exposed tissue in this body of work; the inference from Complex I activity to ATP output, while physiologically standard, was not confirmed with a direct ATP assay in the cited source. No molindone-specific data exist at any tier.
• At-risk subgroups: None identified specific to molindone.
• Counter-evidence on record: None located beyond the general caveat that this is inferred, not directly measured, ATP impact, and that it is a class, not drug-specific, signal.

Regulatory status

Not addressed in the FDA label.

Population evidence

Same study as the Mitochondrial burden entry: Maurer & Möller 1997, human brain cortex homogenate, haloperidol/chlorpromazine/risperidone/zotepine/clozapine tested, molindone not tested. Complex I is the first and rate-limiting entry point of the electron-transport chain into oxidative phosphorylation, so its inhibition mechanistically constrains downstream ATP synthesis, though ATP yield itself was not directly measured in this study.

Mechanism

Named for the class, inferred rather than directly measured for ATP output: Complex I thiol oxidation reduces electron flux into the chain, constraining ATP synthase substrate supply.

Counter-evidence held on record

Do not read this as measured fatigue or low cellular energy in molindone patients. It is an inferred, class-level, laboratory-only signal and does not support CoQ10, carnitine, or riboflavin supplementation (see Fails verification for the CoQ10 claim specifically).

Limits on what this shows

• Why no grade: Dual condition met. The FDA and DailyMed labels address only white-cell effects - "leukopenia, neutropenia and agranulocytosis have been reported temporally related to antipsychotic agents" - which is an immune/bone-marrow idiosyncratic-reaction warning, not an oxygen-delivery finding, and is correctly excluded here rather than mis-filed as a nutrient or oxygen-delivery deficiency (Hard Rule 2). No study measuring hemoglobin, hematocrit, or tissue oxygenation specifically in molindone-treated patients, or in a defensible first-generation-class cohort, was located. No circulating consumer claim alleging molindone-driven oxygen-delivery impairment was found to rebut as white.

Channels tried

This repair pass: the Crossref works API was reachable and returned clean, on-topic-but-empty result sets for "haloperidol methemoglobin oxygen consumption mitochondria oxygen utilization," "haloperidol hemoglobin oxygen methemoglobin erythrocyte," "first generation antipsychotic anemia hemoglobin erythropoiesis," and "neuroleptic antipsychotic mitochondrial oxygen consumption respiration" - none of these surfaced a haloperidol- or other non-phenothiazine-typical-specific study of hemoglobin, methemoglobin, or mitochondrial oxygen consumption (the molindone dossier's own Mitochondrial burden/ATP production source, Maurer & Möller 1997, measured Complex I enzyme inhibition, not oxygen consumption directly, so it was not extended to this category without a source that actually measures oxygen uptake). PubMed, Europe PMC REST, and NCBI eutils/DailyMed were blocked again this pass with the connection itself refused (HTTP 403 policy denial), consistent with this project's documented org-level restriction. This is now a better-confirmed literature gap than the original pass found, with that residual tooling limitation disclosed.