Baclofen

Limits on what this shows

• Key negative/null findings: The MTT readout and protein carbonyl were unchanged in [7]. A PubMed listing for baclofen with mitochondria terms returned 12 records and, judged by title, none reported baclofen-induced mitochondrial damage.
• At-risk subgroups: None identifiable from this evidence. A mitochondrial disease population is not studied with baclofen in any source opened.
• Counter-evidence on record: No study reporting mitochondrial toxicity from baclofen was located. All four studies are animal or in vitro, in disease or stress models, with parenteral or direct exposure; the brown-fat study measured GDP binding, not ATP or oxygen consumption, and its relevance to oral therapeutic dosing in adult humans is untested.

Regulatory status

No mitochondrial warning was identified in the label sources read [1-5].

Population evidence

Tier B, four independent preclinical studies, no human data. (a) Jahani et al. 2025: in mouse brain mitochondria, the MTT viability readout (a succinate-dehydrogenase-based assay) was unchanged after 7 days at 10 mg/kg intraperitoneally; ATP, membrane potential, and complex-specific activity were not measured [7]. (b) Liu et al. 2015, Scientific Reports: rats with chronic cerebral hypoperfusion (permanent bilateral carotid occlusion) given baclofen 12.5 and 25 mg/kg; baclofen suppressed excessive autophagy markers (LC3-II, Beclin 1, atg5, atg7), reduced the surface and mitochondrial expression of connexins 43 and 36 that rose under hypoperfusion, and increased Bcl-2/Bax and reduced cleaved caspase-3 [9]. (c) Primary rat retinal ganglion cells exposed to cobalt chloride (a chemical hypoxia model) in vitro: baclofen reversed apoptosis, lowering BAX and cleaved caspase-3 and raising BCL-2, with Akt activation and PERK-pathway suppression [10]. (d) Rothwell et al. 1985, J Pharm Pharmacol: in anesthetized rats, intravenous baclofen or injection into the ventromedial hypothalamus raised body and rectal temperature and increased GDP binding in brown-fat mitochondria, mediated through the sympathetic nervous system [11]. Doses for [10] and [11] were not retrievable from the extracts read.

Mechanism

Protective findings run through GABA-B signaling to Akt/mTOR and apoptosis regulators ([9], [10]); the brown-fat effect is centrally triggered sympathetic activation ([11]). None is a mitochondrial toxicity mechanism, and none is a nutrient mechanism.

Counter-evidence held on record

This entry records an absence of a harm signal plus protective and thermogenic animal findings. It is not a claim of mitochondrial damage or protection in patients, and it does not support any mitochondria-targeted supplement. No supplement is recommended.

Limits on what this shows

• Key negative/null findings: Protein carbonyl (protein oxidation) and the MTT mitochondrial-viability readout did not change in the Jahani study [7]. Short-term and lower-dose arms are reported in the paper but only the long-term 10 mg/kg effect was extracted here.
• At-risk subgroups: None identifiable: no human cohort exists. Baclofen-treated patients with renal impairment accumulate the drug, but no redox consequence of that has been studied.
• Counter-evidence on record: No study showing baclofen increasing oxidative stress was located. The only limitation is design: mouse and cell work, intraperitoneal dosing, healthy brain in one study and an LPS inflammation model in the other, and no human redox biomarkers (for example urinary isoprostanes or plasma glutathione) in anyone taking oral baclofen.

Regulatory status

Not addressed in the labels or monograph opened [1-5]. These are independent mechanistic literature findings, not labeled effects.

Population evidence

Tier B, no human data. (a) Jahani et al. 2025, Toxicology Reports: male BALB/c mice, baclofen 1, 5, and 10 mg/kg intraperitoneally for 1 day (short-term) or 7 days (long-term), with an 8-day withdrawal arm; in brain mitochondria at 10 mg/kg long-term, reactive oxygen species fell (p < 0.0321), malondialdehyde fell (p < 0.03), and glutathione rose (p < 0.0213); protein carbonyl did not change significantly [7]. (b) Deng et al. 2024, Antioxidants: C57BL/6 mice given baclofen 10 mg/kg intraperitoneally for 5 days before an LPS challenge, plus IPEC-J2 intestinal epithelial cells at 100 micromolar; superoxide dismutase, glutathione, and total antioxidant capacity rose in mouse intestine, catalase rose in cells, and iNOS fell [8]. Effect sizes with confidence intervals were not retrievable from the extracts read.

Mechanism

Deng et al. frame the effect as GABA-B receptor activation modulating the TLR4/MyD88/NLRP3 inflammatory pathway; Jahani et al. report the redox shift without a stated pathway. Both are baclofen's own GABA-B pharmacology, not a nutrient mechanism.

Counter-evidence held on record

This is a reverse-direction, preclinical finding. It is not evidence that baclofen protects anyone against oxidative stress, and it is not a reason to add or avoid any antioxidant. No supplement is recommended.

Limits on what this shows

• Why no grade: The dual condition in Step 7 is met. No labeled or described mechanism connects baclofen to copper absorption, transport, or ceruloplasmin. A Europe PMC title/abstract search for baclofen with copper, ceruloplasmin, or trace element returned 9 records, none reporting an effect of baclofen on copper status: they were Wilson disease and neurodegeneration case reports in which intrathecal baclofen treated dystonia (indication context, not a drug effect), one StatPearls article, an analytical chemistry paper, and a rat hippocampal paper on zinc and GABA-B potentials (receptor electrophysiology). No consumer source located attributes a copper effect to baclofen.

Channels tried

Europe PMC REST search via the page-fetch tool: completed, 9 hits, clean for any copper effect. PubMed search-listing, free-text "baclofen copper OR iron OR ferritin OR anemia": completed but non-specific (iron-deficiency reviews with no baclofen content). PubMed search-listing, "baclofen copper" alone: HTTP 429 rate limit, not retried per the proxy's instruction. Crossref bibliographic search "baclofen copper iron": completed, ten results, all metallurgy standards and general copper/iron references, none about baclofen. DailyMed/FDA labels [1-5]: no copper language in the sections read. OpenAlex search: HTTP 429. Semantic Scholar and Unpaywall: no usable attempt (direct connection refused by organization policy; not retried through the fetch tool). General web search: one query completed with no relevant result; the allowance was then exhausted. Repair pass 2026-10-04: a PubMed field-tagged copper/ceruloplasmin listing was attempted again and returned HTTP 429 (not retried, per the fetch tool's instruction); the web search allowance was confirmed exhausted (200 of 200).

Assessment

A genuine gap with moderate confidence: one completed, targeted Europe PMC search plus one non-specific PubMed and one non-specific Crossref search found nothing, but a clean PubMed title/abstract copper-only listing was not obtained because of rate limiting. Re-check when PubMed access recovers.

Limits on what this shows

• Why no grade: The dual condition in Step 7 is met. No labeled mechanism connects baclofen to iron absorption, transport, or storage, and no label read lists a hematologic adverse reaction. A field-tagged PubMed search (baclofen in title/abstract with iron, ferritin, anemia, or anaemia) returned 25 records; the 10 on page 1 were neurodegeneration-with-brain-iron-accumulation reviews (PKAN, PLA2G6, fatty acid hydroxylase, ISCA2), a Friedreich ataxia review, Lesch-Nyhan, a spasticity versus periodic-limb-movements paper, and a veterinary poisoning review, in which baclofen appears as symptomatic treatment of the indication, not as a cause of an iron change. Repair pass 2026-10-04: pages 2 and 3 (the remaining 15 records) were then opened. Page 2 held a lactoferrin cDNA microarray paper, sleep-related painful erections (baclofen ineffective; iron and selenium also taken without benefit), opisthotonus, a COVID-19 hiccup report (ferritin and LDH elevated by the infection; baclofen given for hiccups), trigeminal neuralgia drug choice, three neurodegeneration-with-brain-iron-accumulation or pantothenate kinase case reports (baclofen as symptomatic treatment, one with deferiprone), a stiff-person syndrome case report (pernicious anemia in the patient; baclofen as treatment), and a paclitaxel-induced sickle cell crisis report (hemolysis from sickle disease; baclofen given for pain). Page 3 held an infant rhesus macaque autoimmune hemolytic anemia report (baclofen one of several drugs listed; causality not established), a G6PD-deficiency priapism report, hiccup, BPAN, PLA2G6, restless legs, infant sleep, and a rat hippocampal zinc and GABA-B potentials paper. Judged by listing titles and extraction notes, none is a study measuring iron, ferritin, or transferrin in baclofen-treated patients; in each, baclofen is symptomatic treatment or one of several co-administered drugs, and the anemia, ferritin, or hemolysis finding is attributed to the underlying disease. Those individual articles were not opened (article pages return no content through the fetch route), so they are not cited and not used as case-report evidence. No consumer source located attributes an iron effect to baclofen.

Channels tried

PubMed search-listing, field-tagged baclofen with iron/ferritin/anemia terms: completed, 25 records, page 1 reviewed. PubMed free-text variants: completed but non-specific. Europe PMC REST iron/hemoglobin query: HTTP 400 on a malformed boolean, then HTTP 429 on repeated attempts, so not completed. Crossref bibliographic "baclofen copper iron": completed, no baclofen hit. OpenAlex search: HTTP 429. DailyMed/FDA labels [1-5]: no iron, ferritin, or hematologic language in the sections read. General web search: attempted, then allowance exhausted before an iron-specific query could run. Repair pass 2026-10-04: PubMed pages 2 and 3 of the field-tagged listing completed (see above, all 25 records now viewed). A Europe PMC iron/ferritin/transferrin title/abstract query returned HTTP 429 and was not retried. Sun et al. 2020 (PMID 32651311), previously not opened, was opened as an OpenAlex single-record abstract: a baclofen rat Alzheimer's model reporting oxidative stress markers (MDA, SOD, GSH-Px, CAT) and PI3K/Akt signaling; iron was not addressed, so it does not bear on this entry.

Assessment

A genuine gap with higher confidence than before: all 25 records of the field-tagged PubMed listing are now viewed and none measures an iron effect of baclofen. Residual tooling caveat: the Europe PMC iron query did not complete and article-level pages could not be opened. Re-check when access recovers.

Limits on what this shows

• Key negative/null findings: A PubMed listing for baclofen with gut microbiota terms returned two records: this study and a 2018 rodent microbiome study in compulsive alcohol seeking (Jadhav et al., PMID 30172845) that was not opened and is not relied on. No human fecal-microbiome study in baclofen users was located.
• At-risk subgroups: None identifiable from this evidence.
• Counter-evidence on record: No finding of baclofen-induced dysbiosis was located in the PubMed listing, the Crossref bibliographic search ("baclofen gut microbiome microbiota": ten results, none about baclofen), or the web search. The study was done in an LPS-challenge model, not in untreated animals or patients, so it does not show the effect of baclofen on a baseline human microbiome.

Regulatory status

No microbiome language in the label sources read [1-5]. The monograph lists nausea (very common) and constipation, vomiting, and diarrhea (common) as adverse reactions; these are symptoms, not microbiome measurements [5].

Population evidence

Tier B, single study, no human data. Deng et al. 2024, Antioxidants: adult male C57BL/6 mice (6 to 8 weeks), baclofen 10 mg/kg intraperitoneally for 5 consecutive days before LPS. LPS alone reduced the ACE, Chao1, Shannon, and Simpson alpha-diversity indices; baclofen pretreatment "had the potential to counteract" this; Lactobacillus rose substantially and Enterobacteriaceae fell [8]. Sample sizes and confidence intervals were not retrievable from the extract read.

Mechanism

Not established. The authors tie GABA-B activation to reduced oxidative stress and TLR4/MyD88/NLRP3 signaling, with the microbial shift reported alongside; causality between the two is not shown by the extract read.

Counter-evidence held on record

This is a single mouse-model finding in the protective direction. It is not a reason to take probiotics or to expect a microbiome benefit from baclofen. Constipation and diarrhea on the label are adverse reactions to be reported to a prescriber, not microbiome findings.

Limits on what this shows

• Why no grade: The dual condition in Step 7 is met on what could be searched. The labels and monograph read list no hematologic abnormality (labeled lab changes are AST, alkaline phosphatase, and blood sugar). The one PubMed free-text search for baclofen with hemoglobin, anemia, hematocrit, or oxygen saturation returned ten records, none about baclofen. The iron-field-tagged search (see Iron) also surfaced no baclofen-related anemia report on page 1. The drug's own CNS depression (labeled as additive with alcohol and other CNS depressants) is intended-class or toxicity pharmacology, not a nutrient-mediated oxygen-delivery deficiency, and is not graded here (Hard Rule 2). No consumer source located attributes an oxygen-carrying-capacity effect to baclofen.

Channels tried

PubMed free-text listing: completed, ten non-baclofen records. PubMed field-tagged iron/anemia listing: completed, page 1 only. Crossref bibliographic "baclofen hemoglobin anemia oxygen": HTTP 429, not completed. Europe PMC: HTTP 400 then 429, not completed. OpenAlex search: HTTP 429. DailyMed/FDA labels [1-5]: read, no hematologic language. General web search: allowance exhausted before an oxygen-specific query could run. Semantic Scholar and Unpaywall: no usable attempt. Repair pass 2026-10-04: a Europe PMC title/abstract query for baclofen with hemoglobin, anemia, hemolysis, or methemoglobin completed (six records: an esophageal Dieulafoy lesion bleeding case, a dog baclofen toxicosis lipid-emulsion paper, a paclitaxel sickle cell crisis report, a stiff-person syndrome case report, a G6PD-deficiency priapism report, and a 1993 rat cerebellum nitric oxide paper); none reports baclofen causing or altering hemoglobin, red-cell mass, hemolysis, or methemoglobin, and the PubMed iron/anemia pages 2 and 3 added only reports in which baclofen was symptomatic treatment or one of several co-administered drugs. A PubMed hemolysis/methemoglobin/oxygen-consumption/hypoxemia listing returned only the PubMed homepage (no usable records), and Europe PMC queries for hypoxemia, oxygen consumption, and respiratory depression returned HTTP 400 or 429 and were not retried; a web search for baclofen overdose and respiratory depression was refused because the allowance was exhausted. Baclofen-related respiratory depression or hypoxemia in overdose, if located, would be the drug's own toxicity and would be filed as such under Hard Rule 2, not graded as an oxygen-delivery finding.

Assessment

A probable gap, flagged as partly tooling-limited rather than cleanly exhausted. The hematology half (hemoglobin, anemia, hemolysis) is now supported by one completed Europe PMC query and the full iron/anemia PubMed listing with no baclofen effect found; the tissue-oxygenation and respiratory half could not be searched. Originally only two PubMed listings had completed and the secondary hosts were rate-limited. Re-check hemoglobin, hematocrit, and respiratory-depression literature when tooling recovers.

Limits on what this shows

• Why no grade: The dual condition in Step 7 is met on what could be searched. The Jahani 2025 study explicitly did not measure ATP, membrane potential, or complex-specific activity [7]. The Rothwell 1985 brown-fat finding (GDP binding) [11] and the cell-survival findings [9, 10] are used under Mitochondrial burden and are deliberately not carried over here as if they were ATP evidence, because none measured ATP synthesis or energy output. Fatigue (2 to 4 percent) and weakness (5 to 15 percent) appear as labeled adverse reactions [1]; these are clinical outcomes of a muscle relaxant, not measured ATP deficits (Hard Rule 2). No consumer source located claims baclofen lowers ATP or cellular energy.

Channels tried

PubMed field-tagged listing, baclofen with ATP, energy metabolism, or respiratory chain: completed, 51 records, page 1 reviewed (K-ATP channel electrophysiology, Leigh syndrome GeneReviews, GABA-B neuron papers; none reporting baclofen effects on ATP synthesis). PubMed free-text "baclofen mitochondria": completed, 12 records. Crossref bibliographic "baclofen mitochondrial ATP production": completed, six results, none about baclofen. PubMed ATP/oxidative-phosphorylation free-text variant: HTTP 429. Europe PMC mitochondria/ATP: HTTP 400 then 429. OpenAlex search: HTTP 429 (single-record lookup by PMID did work and was used for the Rothwell abstract). Web search: one mitochondria/oxidative-stress query completed; later ATP-specific queries refused because the allowance was exhausted. Repair pass 2026-10-04: page 2 of the 51-record PubMed listing was opened and held ten records, nine of them GABA-B or K-ATP channel electrophysiology and pharmacology papers (for example sulfonylurea-sensitive K+ current in dopamine neurons, a K-ATP channel involvement in the antidepressant-like effect of baclofen in mice, the carotid body GABA-B and TASK-1 paper, GABA-B modulation of insulin secretion) plus a muscle cramps in cirrhosis systematic review; none reports baclofen altering ATP levels, ATP synthesis, oxygen consumption, or respiration. K-ATP channels are metabolic sensors and their involvement is a signaling observation, not a measurement of ATP production, so it is not graded. A Europe PMC ATP/respiration query was rejected as too long, a shorter oxygen-consumption query returned HTTP 429, and two PubMed free-text ATP/respiration variants returned the homepage or HTTP 404. The Rothwell 1985 brown-fat GDP binding result [11] was reconsidered for a yellow mechanistic grade and again not used: GDP binding is a marker of uncoupling-protein activation, no ATP or respiration was measured, and carrying it here would double count the Mitochondrial burden entry (see Overlap map).

Assessment

A probable gap, flagged as partly tooling-limited: pages 3 to 6 of the 51 PubMed records were not reviewed (page 2 now was) and Europe PMC could not be completed. Re-check when access recovers.