Montelukast

Limits on what this shows

• Null findings: None in humans (not studied). Mouse and cell results are model-specific and abstract-level.
• At-risk subgroups: Children are proposed as more susceptible to the neuronal effects because of brain maturation stage [7]. This is the authors' hypothesis, not a measured subgroup finding.

Regulatory status

The label states the neuropsychiatric mechanisms are "not well understood" and does not mention mitochondria [1].

Population evidence

No human study measured mitochondrial function with montelukast. Tier B only: Marques 2022 [7] (embryonic chicken neuron-enriched cells and mice, metabolomics plus proteomics): 18 new montelukast metabolites identified, reaction with glutathione confirmed, interference with brain glutathione detoxification, neurotransmitter and neurosteroid (HPA axis) pathways and neuronal mitochondrial function; concentrations and specific mitochondrial parameters were not in the abstract (repair pass: the OpenAlex abstract confirms the same wording and no ATP term; the full text is open access but the publisher page and PDF were blocked by robots exclusion, so it was not read [26]). Wang 2019 [6] (human bronchial epithelial Beas-2b cells): increased PGC-1 alpha, NRF-1, TFAM, mitochondrial mass, mtDNA/nDNA ratio and cytochrome B via cAMP/CREB (beneficial). Protective in rodent or hepatocyte toxin models: prednisolone liver injury, arsenic liver injury, quinolinic/malonic acid neurotoxicity, pemetrexed hepatocyte cytotoxicity (PMID 35291928; listing-level summary: reduced mitochondrial ROS among antioxidant effects, ATP not mentioned; the PMC full text returned a bot check) [27], acetaminophen (JC-1 mitochondrial membrane potential restored in primary hepatocytes; on re-opening the Frontiers PDF, the text describes this qualitatively at 1, 5 and 10 uM montelukast with 2.5 mM acetaminophen, with no numerical quantification and no ATP or oxygen-consumption measurement) [8][9].

Mechanism

Present but contested. Harm pathway: glutathione conjugation and neuronal mitochondrial interference [7]. Protective pathway: cAMP/CREB/PGC-1 alpha biogenesis [6], and rescue of toxin-induced mitochondrial dysfunction [8][9]. Doses in animal studies are not shown to match human exposure.

Counter-evidence held on record

Counter-evidence and confounding The same drug is protective in most models. Mouse brain exposure and neuronal cell dosing were not compared with human plasma levels. The neuropsychiatric boxed warning is a drug toxicity and is not a nutrient deficiency; it is cited only as the clinical context for [7]. Framing caution Preclinical hypothesis. Do not state that montelukast damages human mitochondria, and do not recommend any mitochondrial supplement from this entry.

Limits on what this shows

• Null findings: Human oxidant and antioxidant status did not normalize on montelukast [4][5].
• At-risk subgroups: None identified. Asthmatic children have higher oxidant load regardless of treatment [4].

Regulatory status

Not mentioned in the label [1].

Population evidence

Dilek 2015 [4]: three groups, 31 newly diagnosed untreated asthmatic children, 32 treated with montelukast for at least 6 months, 40 healthy controls. Total oxidant status 13.1 +/- 4 (untreated), 11.1 +/- 4.1 (montelukast), 6.3 +/- 3.5 (controls) umol H2O2 equiv/L; total antioxidant status 1.4 +/- 0.5, 1.5 +/- 0.5, 2.7 +/- 0.6 mmol Trolox equiv/L. Both asthma groups differed from controls (p less than 0.05); the montelukast effect was minor and not significant. DNA damage (comet assay) 18.2, 16.7 and 13.7 a.u.; only the untreated group differed from controls. Dilek 2016 [5] (60 asthmatic children, 35 controls; montelukast monotherapy at least 4 months): plasma total thiol 0.20 mmol/L in the montelukast group versus 0.21 in controls (not different) and 0.18 in untreated asthma (p less than 0.001 versus controls); total antioxidant status stayed below control in both asthma groups. Abstract-level reads; cross-sectional comparison, not randomized. Repair pass: Turk J Med Sci 2007 [28], 25 atopic asthmatic children given a beta-agonist for one month and then montelukast 5 mg for one month, with 25 nonatopic controls (before-and-after, not randomized): serum paraoxonase activity rose from 211.0 +/- 123.2 to 323.2 +/- 189.7 U/L (significant; control 230.5 +/- 108.7), while the lipid peroxidation marker malondialdehyde moved from 4.2 +/- 0.2 to 4.7 +/- 0.6 nmol/mL (not significant; control 4.6 +/- 0.4). Read through a summarizing tool from the journal PDF, so values are as reported there and were not independently checked. This adds a third human cohort with montelukast in it; direction is neutral to favorable, which keeps the entry yellow in the reverse direction.

Mechanism

Present, reverse direction: CysLT1 blockade with downstream KLF2 activation reduced oxidative stress and apoptosis in IL-1 beta-treated chondrocyte-like cells (abstract-level listing summary) [10]; lower H2O2 and TBARS and a higher GSH/GSSG ratio with acetaminophen in mice [9]. A harm-direction mouse study (young ICR mice, 28 days) reported depressive-like behaviour with hippocampal oxidative stress and raised inflammatory and stress-hormone markers, abstract only [10]. The glutathione reactivity in [7] is a possible harm mechanism in neurons.

Counter-evidence held on record

Counter-evidence and confounding The markers are disease-linked: untreated asthma already differs from controls. Treatment groups were not randomized and were treated for different durations. Animal studies use toxin models. Framing caution This is not an antioxidant-depletion claim. Do not present montelukast as an antioxidant therapy, and do not recommend antioxidant supplements from this entry.

Limits on what this shows

• Null findings: No human iron study found. The in silico speciation model found no effect on iron distribution [33]; it is a model, not a measurement.
• At-risk subgroups: None identified.

Regulatory status

Not mentioned in the label [1].

Population evidence

Mouse study (MPTP model, male C57BL/6; PMID 41712018, Neuromolecular Med 2026), abstract-level, via PubMed listing summary: montelukast reduced iron deposition and ferroptosis markers along the GPX4/ACSL4/5-LOX pathway [19]. Sample size not stated. Repair pass, Springer record opened (DOI 10.1007/s12017-025-08895-5) [32]: the abstract says iron deposition and malondialdehyde fell and glutathione and superoxide dismutase activity were restored; no numbers, assay method, dose or group size are in the abstract, and the Methods and Results were not accessible. Computational plasma speciation model, montelukast-specific (conference abstract, 10 metals modeled, iron among those named): no discernible effect of montelukast on iron distribution, zinc consistently decreased except one complex, copper not named [33]. Human: an uncontrolled observational study (n=578) of montelukast plus acetylcysteine in COVID-19 pneumonia reported ferritin normalizing by day 20; the combination and acute illness make it unusable for a montelukast iron effect [19]. A hemodialysis pruritus crossover trial (n=16) recorded one anemia withdrawal and one hemoglobin drop in the placebo period, which is a safety note and not iron evidence [20].

Mechanism

Anti-ferroptotic signaling in a toxin model. No systemic iron absorption, binding or loss mechanism for montelukast was found.

Counter-evidence held on record

Counter-evidence and confounding A neuro-model result does not speak to systemic iron status. Abstract not read beyond the listing summary. Iron deficiency in children with asthma, if present, is confounded by atopy, diet and disease. Framing caution Not an iron-deficiency or iron-overload claim. Do not recommend iron supplementation or iron avoidance from this entry.

Limits on what this shows

• Why no grade: Dual condition met: no Tier B or C source for copper with montelukast, and no circulating copper claim to rebut. This is a disclosed gap, not a finding of safety.

What was searched

PubMed listing "montelukast AND (iron OR ferritin OR copper OR ceruloplasmin OR zinc)" (22 records; 15 visible in the summarized read): no human copper measurement in montelukast users; hits were pruritus reviews, formulation chemistry (copper oxide nanoformulation, mice), and a ferroptosis mouse study. WebSearch "montelukast vitamin B12 OR iron OR copper deficiency asthma patients" returned titles only (trace-element asthma studies without montelukast data). A Tier D depletion page for montelukast names no nutrient [23]. Disease context opened: a children's asthma trace-element paper reports higher copper and ceruloplasmin and lower zinc than controls, with no montelukast stratification and no sample size on the page [18], so it is a disease association and does not support a drug effect. Repair pass: WebSearch "montelukast copper complex chelation OR copper levels OR serum copper asthmatic children before and after montelukast" returned only disease-association papers. A meta-analysis of serum copper in childhood asthma (11 studies, 1,006 participants) found no pooled difference from controls (SMD -0.032, 95% CI -0.291 to 0.228, I2 73.5%) and no montelukast stratification (abstract only) [34]; an Egyptian case-control study (30 asthmatic children, 15 controls) reported higher serum copper in asthma (122 +/- 31.5 versus 103.3 +/- 21.1 ug/dL) with no named medication [34]. The montelukast plasma speciation model does not name copper [33]. None of these has copper measured in montelukast users against non-users.

Channel outcomes

DailyMed label: opened, no copper content. PubMed listings: opened intermittently (several 429s); the repair pass added a computational model abstract and its PDF (copper not named), two copper-in-asthma disease papers, and further WebSearch phrasings (ceruloplasmin, serum copper, chelation). OpenAlex, Crossref: some single-record calls blocked by 429. Mixed result: genuine gap in the opened literature, with some tooling limitation (seven records in the original listing were not visible).

Limits on what this shows

• Null findings: Little overall community change with montelukast in the long COVID study [13]. Oral montelukast is not an antibiotic and no C. difficile or dysbiosis signal was found in the sources read.
• At-risk subgroups: None identified.

Regulatory status

Diarrhea and abdominal pain are among the most common trial reactions; the label does not mention the microbiome [1].

Population evidence

Fan 2026 [11], single-center RCT, 100 children aged 6 to 11 with asthma, montelukast plus inhaled corticosteroid versus inhaled corticosteroid alone, 12 weeks, 16S nasal and fecal sequencing, SCFA by LC-MS/MS. The combination arm had greater clinical improvement, higher nasal and fecal alpha diversity, fewer Moraxella and Haemophilus, more Faecalibacterium and other SCFA producers, higher serum SCFA, regulatory T cell and IL-10 responses, and lower Th2 cytokines; SCFAs partially mediated the FeNO improvement. Repair pass, abstract read from the Hungarian Academy repository record [29]: 50 children per arm; combination versus ICS alone reduced FeNO by an interaction beta of -10.24 ppb (95% CI -16.37 to -4.11, P = 0.001), with C-ACT beta +1.83 and FEV1% beta +1.87 (both P below 0.05); SCFAs mediated 30.0% (total SCFA) and 37.5% (butyrate) of the FeNO improvement; authors report a comparable safety profile. Alpha diversity values, blinding and the SCFA concentrations were not retrieved (publisher page 403, repository PDF 401). Italian pilot [12]: 20 children aged 6 to 9, no healthy controls, 16S; antileukotriene use (about 50% of children) correlated with predicted functional groups: LPS-producing r=0.50, mucosa-associated r=0.46, sulfate-reducing r=0.40, butyrate-producing r=0.27; hypothesis-generating, use recorded only as present or absent, montelukast not named. Long COVID proof-of-concept [13]: short-term montelukast mainly enriched Dialister with little overall community change (abstract-level, listing only). Herbal decoction RCT [13], n=116, montelukast active comparator in children with post-infectious cough (abstract-level).

Mechanism

Not established. Authors propose reduced Th2 inflammation shifting airway and gut communities; the add-on design cannot separate montelukast from inhaled-steroid effects.

Counter-evidence held on record

Counter-evidence and confounding Comparison is add-on versus no add-on; open versus blinded status was not retrieved. Asthma itself is associated with altered gut communities. No study tested montelukast against non-use in healthy people. The 20-child pilot is cross-sectional and cannot give direction. Framing caution This is a reverse-direction finding, not a microbiome-harm claim and not a probiotic recommendation.

Limits on what this shows

• Why no grade: Dual condition met: no Tier B or C evidence of hematologic harm, and no circulating claim found. Kept separate from ATP production. This is a disclosed gap, not a finding of safety.

What was searched

PubMed listing "montelukast hemoglobin" (9 records, abstract-level): a hemodialysis pruritus crossover trial (n=16) with one hemoglobin decrease in the placebo period and one anemia withdrawal (arm unspecified); HbA1c trials (glycemic marker, not oxygen carrying); case reports of eosinophilic gastroenteritis with bleeding in which authors did not implicate montelukast; no methemoglobin or hemolysis signal [20]. WebSearch "montelukast hemoglobin anemia OR methemoglobin OR hemolysis adverse reaction": titles only, including a sickle cell trial record (NCT01960413) whose page returned no content. The label and StatPearls list no anemia or hemolysis [1][2]. Repair pass: (a) a double-blind placebo-controlled crossover trial in 14 non-asthmatic highly trained male endurance athletes (VO2max above 70 ml/kg/min), single 10 mg dose, oxygen uptake and ventilation measured during exercise at -15 C: no effect on performance variables, time to exhaustion 332.3 +/- 45.8 s on montelukast versus 340.1 +/- 53.3 s on placebo (P = 0.22) [35]; hemoglobin was not reported, and a single dose in healthy athletes says nothing about chronic use; (b) the phase 2 randomized trial of montelukast in sickle cell disease (42 randomized, 8 weeks, Blood Adv 2020) lists hemoglobin and reticulocytes among protocol labs but the paper reports only baseline median values (hemoglobin 9.1 versus 8.7 g/dL, reticulocytes 9.15% versus 8.85%, montelukast versus placebo) and no week 8 hematology [36]; (c) a single pediatric case report of ecchymosis on montelukast gave hemoglobin 12.1 g/dL and hematocrit 34.8% [37]; (d) WebSearch for methemoglobinemia, hemolytic anemia or G6PD with montelukast returned only reports not involving montelukast (opened: a mothball hemolysis case and a levofloxacin case). No hemoglobin or oxygen-carrying change attributable to montelukast was found, and none of these is a nutrient finding.

Channel outcomes

PubMed listing: opened after two 429s. ClinicalTrials.gov: opened, empty. Mixed: genuine gap for a drug-caused effect; the sickle cell trial record was recovered in the repair pass but reports no week 8 hematology. Montelukast's intended airway effect (bronchoconstriction relief) is a pharmacologic effect and not an oxygen-delivery nutrient finding (Hard Rule 2).

Limits on what this shows

• Why no grade: Dual condition met for ATP specifically: no Tier B or C source measuring ATP was opened, and no circulating "low energy" claim for montelukast was found. Shared mitochondrial evidence is recorded in the Overlap map. This is a disclosed gap, not a finding of safety.

What was searched

PubMed listing "montelukast AND (mitochondrial OR mitochondria)" (17 records read at listing level): none reported ATP synthesis, ATP content or oxygen consumption; biogenesis markers [6] and membrane potential [9] are structure and potential, not ATP output. WebSearch "montelukast ATP production OR ATP levels OR bioenergetics OR oxidative phosphorylation": no relevant result (a sepsis trial title and a leukemia OXPHOS paper unrelated to montelukast). The two follow-up PubMed ATP listing queries returned 429 and were not read in the original pass. Repair pass: PubMed listing "montelukast AND (ATP OR oxidative phosphorylation OR oxygen consumption)" was read (18 records read at listing level; two further PMIDs, 19748977 and 19484671, were cut off by the page and not read): ATP appears only as an energy source for transporter assays, as K(ATP) channel pharmacology, or as a ligand, and none measured cellular ATP content, ATP synthesis or respiration with montelukast [38]. Opened with no ATP result: the acetaminophen mouse paper PDF (ATP and oxygen consumption do not appear) [9], the Wang 2019 abstract (mitochondrial mass, mtDNA/nDNA, cytochrome B and PGC-1 alpha, no ATP term) [6], the pemetrexed hepatocyte listing (ATP not mentioned) [27]. Two glucose/galactose mitochondrial toxicity assay papers (ATP readout) were opened at abstract level; neither mentions montelukast. The Marques 2022 review-style preprint by the same authors reports no mitochondrial data for the drug [26].

Channel outcomes

Tooling-limited in part: the ATP-specific PubMed listings were rate limited, so the 2022 pemetrexed hepatocyte paper (PMC9208499, bot check again on 2026-10-08) could not be checked in full for an ATP assay; its listing-level summary does not mention ATP. Re-check the PMC full texts of PMC9208499 and Wang 2019 when access is restored.